Isotopic paleodiet analysis at 2nd c AD Rogowo, PL

Laurie J. Reitsema and Tomasz Kozłowski

Anthropological Review • Vol. 76 (1), 1–22 (2013)

Diet and society in before the state: stable isotope evidence from a Wielbark population (2nd c. AD)

Laurie J. Reitsema1, Tomasz Kozłowski2

1Department of Anthropology, University of Georgia, 250 Baldwin Hall, Jackson St., Athens, Georgia, 30602, USA 2 Department of Anthropology, Nicolaus Copernicus University, ul. Lwowska 1, 87-100, Toruń, Poland

Abstract: The 1st–4th c. AD of Eastern Europe is relatively understudied bioarchaeolog- ically due to the fragmentary nature of its cemeteries. Here, we report the first stable isotope analysis of Wielbark diet using stable carbon and nitrogen isotope signatures from both collagen and carbonate of 30 individuals from Rogowo, a 2nd c. Wielbark cemetery in North-Central Poland. Diet at Rogowo was pri- marily based on terrestrial foods and included millet, a C4 plant cultivated by many Slavic populations in Europe. Anadromous fish likely supplemented the diet, which is clarified when considering collagen and carbonate data in tandem. Stable isotope differences between the sexes indicate that men and women may have consumed different foods, although there is a possibility that women immigrated to Rogowo from an isotopically different region of Europe. No significant differences are noted in δ13C or δ15N of women with and without grave goods, suggesting little social differentiation within the Wielbark culture, at least in terms of daily food access. Reconstructing human diet in Europe through stable isotope analysis is problematic because of the relative isotopic homogeneity in this region of the world. This study further demonstrates the utility of using both carbonate and collagen stable isotope data in tandem to reconstruct past European diet.

Key words: paleodiet, Roman era, Wielbark culture, isotopes, Poland

Introduction riod of Roman influence or the Roman The archaeological culture known era by archaeologists, although the ge- as the Wielbark occupied regions of ographic regions of modern-day Poland North-Central Europe between the 1st were never incorporated into the Roman and 4th centuries AD (Barford 2001). Empire. Initially settled in the areas sur- This time period is referred to as the pe- rounding the River in Pomera-

Original Article: Received September 13, 2013; Accepted for publication October 21, 2013 DOI: 110.2478/anre-2013-0010 © 2013 Polish Anthropological Society 2 Laurie J. Reitsema and Tomasz Kozłowski nia (Wołgiewicz 1981, Godłowski and Two archaeological trademarks of the Kozłowski 1985), the Wielbark culture Wielbark culture are bi-ritual cemeteries expanded during the first half of the nd2 that include both inhumation and crema- c. AD to Southeast Poland and Wielko- tion burials, and a characteristic lack of polska (Kozak-Zychman 1996). Origins weapons and tools as grave goods (Go- of the Wielbark culture are uncertain. It dłowski and Kozłowski 1985). ­Wielbark may have developed indigenously from settlements are generally open and un- the culture in and around the fortified. A degree of social differenti- Vistula River basin (e.g., Godłowski ation is manifest in the facts that some and Kozłowski 1985; Dąbrowski 2006), graves are whereas some but some believe it may represent im- are inhumations, and that ornaments migrant populations of Germanic and jewelry grave goods are present in and from (Czeka- burials. However, in general, Wielbark nowski 1955; Godłowski and Kozłowski populations appear to have been rela- 1985; Heather 1996; Kozak-Zychman tively egalitarian (Gieysztor et al. 1979). 1996; Wołgiewicz 1981). The issue of Although linked to the Roman world by Wielbark origins and relationships with trade, Slavic populations in Antiquity other Roman era European populations are scarcely described in written records; remains unresolved. After their appear- thus, it falls to archaeology to illuminate ance in North-Central Poland, Wielbark the daily lives of modern-day Poland’s settlements eventually shift further populations in the early centuries AD. south, replacing the In terms of the subsistence econo- in Southern Poland, and apparently de- my, wheat, millet, rye and barley were veloping into the Černjachov culture in the most commonly grown crops by the modern-day Ukraine and surrounding Wielbark culture (Pyrgała 1975). Cattle areas. The relationship between the were the most common domesticated Wielbark culture and later, medieval animal in Roman era Poland, comprising Slavic populations from the and approximately 50 to 80% of the total ar- the Vistula rivers basins is also incom- chaeozoological assemblages (Gładykow- pletely defined (c.f., Kozak-Zychman ska-Rzeczycka et al. 1997; Makowiecki 1996; Dąbrowski 2006). Anthropologi- 2006; Reitsema et al. 2013). Land was cal examinations of skull and tooth mor- tilled on a household level but was prob- phology suggest biological continuity ably collectively owned during the tribal between Antiquity and the early medie- period (Gieysztor et al. 1979). For this val period (Piontek et al. 2008). Recent reason, little within-group dietary differ- preliminary aDNA analyses of a small entiation based on status may be expect- number of and medieval skel- ed at this time. etons from Poland suggest genetic dis- Because most Wielbark cemeteries continuity between these times (Juras comprise cremains and fragmented skel- 2012). Although inconclusive thus far, etal remains, only a limited number of bi- contributions from biological anthro- oarchaeological investigations contribute pology, such as these recent examples, to our understanding of these peoples are desirable to clarify the Wielbark cul- (e.g., Dąbrowski 2006; Kozak-Zychman ture’s relationship to other European 1996, Krenz-Niedbala and Kozłowski populations. 2013; Piontek et al. 2006; Segeda et al. Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 3

2007; Smrćka et al. 2000). Here we pres- tios (δ15N) are reported according to the ent the first stable isotope evidence of equation

Wielbark diet from Rogowo, a cemetery [δ=(Rsample – Rstandard)/Rstandard x1000]. with unusually complete skeletal materi- Stable carbon and nitrogen isotope als, in Kujavian- ratios provide differing information of North-Central Poland. In addition to about human diet. δ13C variation in the reconstructing diet of this Wielbark pop- plant kingdom exists due to differences ulation, we consider relationships be- in photosynthetic pathways used to me- tween diet and sex, age, and presence/ tabolize atmospheric carbon (Park and absence of grave goods to better under- Epstein 1960; Smith and Epstein 1971). stand the dual influences of cultural and δ13C ratios in human tissues therefore biological variables in influencing diet of reflect the local ecosystem, identifying this tribal population. which types of plants were consumed and the influence of environmental- pa 13 Stable Isotope Background rameters on plant δ C values (DeNiro and ­Epstein 1978; Heaton 1999). The Analysis of stable carbon and nitrogen δ13C values of fish often (though not al- isotopes from human bones is an effec- ways) fall outside the typical ranges of tive tool in archaeology for reconstruct- terrestrial foods, and can in many cases ing individual diets of past populations be identified isotopically in diet as a re- (c.f., Katzenberg 2008; Schoeninger sult (Chisholm et al. 1982; Dufour et al. 2011). Plants and animals exhibit sys- 1999; Grupe et al. 1999; Schoeninger and tematic isotopic variation due to differ- DeNiro 1984). ences in their environments, physiolo- Stable carbon isotope ratios can be gies and diets (Ambrose 2000; Park and measured from either bone collagen or δ13 Epstein 1961; Schoeninger and Moore apatite (referred to henceforth as Ccoll δ13 1992; van Klinken et al. 2000). Stable and Cap, respectively). Because collagen isotope signatures are incorporated is formed from amino acids, it primari- into consumer tissues, including bones, ly reflects protein in the diet (­Ambrose which come to represent a composite of and Norr 1993; Tieszen and Fagre 1993). the stable isotope signatures of foods Apa­tite is formed from dissolved bicarbo- eaten in life. By measuring stable iso- nate in the bloodstream which contains tope ratios of ancient bones and com- carbon from all macronutrients; thus, paring the signatures to known ratios carbonate from apatite better reflects en- of various food types, it is possible to ergy sources in diet (i.e., carbohydrates). estimate the types of foods a person typ- Collagen and carbonate provide different ically consumed on a daily basis. Stable and complementary information about isotope values are expressed as a permil foods consumed and when used togeth- (‰) ratio of one of an element’s iso- er, have been shown to be highly effec- topes to another in relation to a stand- tive for reconstructing diet (e.g., Kellner ard of known abundance: Vienna Pee and Schoeninger 2007). Dee Belemnite (VPDB) for δ13C and the The δ13C values of consumers are not Ambient Inhalable Reservoir (AIR) for identical to the values of foods. Rather, δ15N. Both stable carbon isotope ratios there are systematic diet-tissue spacing (δ13C) and stable nitrogen isotope ra- relationships that can be considered in 4 Laurie J. Reitsema and Tomasz Kozłowski order to backtrack from the stable iso- the use of manure or other methods used tope ratios of a consumer to those of to improve soils (ploughing; burning) on δ13 its foods. In general, collagen Ccoll is which domestic animals were grazed. For approximately 5‰ higher in consum- δ13C, the wild/domestic animal difference ers than in the bulk protein diet (van likely represents variations in canopy der Merwe and Vogel 1978). The rela- cover on forage lands, with forest plants δ13 tionship between Cap and diet is less and animals exhibiting lower values due well-established. Controlled laboratory to the “canopy effect” (for more informa- experiments reveal the diet-to-apatite tion on the canopy effect in Europe, see space for δ13C to be approximately 9.5‰ Bocherens and Drucker (2003)). among small mammals (Ambrose and The value of stable isotope analysis in Norr 1993; Tieszen and Fagre 1993), but historic archaeology is that, unlike writ- a review by Garvie-Lok (2001) of pub- ten records, stable isotope signatures lished studies suggests a larger value of are unaffected by record biases in the approximately 12.0‰ may be more ap- prevailing sociopolitical milieu, making propriate for humans. them more direct indicators of an indi- Stable nitrogen isotopes reflect the vidual’s diet. Historical documentation trophic level of a consumer, with carni- of the Roman era in vores systematically exhibiting higher is limited by the fact that most inhabit- δ15N signatures than herbivores due to ants and travelers were illiterate, leav- physiological and chemical processes ing many populations fragmentarily or during metabolism (DeNiro and Epstein subjectively represented in records. In- 1981). The diet-tissue space for δ15N is terpretations of archaeolozoological and generally on the order of 3–5‰ (­Drucker paeobotanical data are limited in that and Bocherens 2004). δ15N values in col- they represent a bulk average of overall lagen can thus be used to estimate the diet among the entire population, rath- amounts and types of animal products er than differential food access among consumed, with higher values in consum- sub-populations. Stable isotope analysis ers suggesting greater animal protein in- has proven invaluable in anthropology take. However, δ15N values also vary with not just for understanding subsistence respect to local climate (Ambrose 1991; behaviors, but also for recognizing sex-, Drucker et al. 2003), soil nitrogen com- age- and status-based differences in food position (Britton et al. 2008; Grogan et access (Katzenberg 1993; Larsen 1997; al. 2000), production systems (Bogaard Polet and Katzenberg 2003; Privat et al. et al. 2007; Commisso and Nelson 2008) 2002; Reitsema et al. 2010; Reitsema and and consumer physiology (Ambrose Vercellotti 2012). 1991; Fuller et al. 2005; ­Hobson et al. 1993), complicating their interpretation. Materials and Methods Isotopic evidence from animal bones in this region of Poland dating to the medi- Rogowo is a Roman era cemetery discov- eval period shows a 3.3‰ difference in ered in 1984 that dates to the 2nd c. AD δ15N values and a 1.4‰ difference in δ13C (Fig. 1). It was excavated in 1999 and values of wild and domestic animals (Re- 2000, and is associated with a nearby set- itsema et al. 2013). Higher δ15N values tlement that covers approximately 6 ha among domestic animals likely reflects (Krenz-Niedbała and Kozłowski 2013). Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 5

The cemetery includes 137 inhumation sampled. No animal bones were found and 151 burials (Chudziak at the site, and so an interpretive faunal 2000). Skeletons were interred in a flexed baseline was established using animals position, many with copper-alloy jewelry from Denmark which were contempora- that has left green stains on skeletal ele- neous with Rogowo, previously reported ments (Fig. 2). One of Eastern Europe’s by Jørkov et al. (2010), and animals from most important trade routes, the Amber medieval Poland, reported here and in Road, connected Rogowo with the Baltic more detail by Reitsema et al. (2013). Sea to the North, and the Roman Empire Collagen was purified from bone us- to the South (Buko 2008). Interregional ing a modified Longin method (Ambro- trade and iron smelting, in addition to se 1990; Ambrose et al. 1997). Briefly, animal husbandry and agriculture, struc- approximately 0.25–0.50 g of bone was tured the economy of this Wielbark pop- crushed to a coarse powder, demineral- ulation (Piontek et al. 2006). ized for 3 to 7 days in 0.2 M HCl, soaked From the cemetery, small pieces of rib in 0.125 M NaOH for 20 hours to remove approximately 1–2 cm in length from 15 other contaminants, solubilized at 90ºC adult males and 15 adult females were in HCl of pH=3, filtered and freeze dried. Three human samples were analyzed in duplicate, and one sample was both pre- pared and analyzed in duplicate. These repeat analyses show replicability to be within 0.1‰ for δ15N and 0.2‰ for δ13C. Of the 30 human samples analyzed for collagen isotope data, 29 were also analyzed for carbonate data. Carbonate was prepared according to protocol de- tailed in Garvie-Lok et al. (2004). Bone was powdered in a steel mortar and pes- tle, deproteinated in bleach for 48 hours, soaked in 0.1 M acetic acid for 4 hours to remove more soluble adsorbed car- bonates, dried, and homogenized in an Fig. 1. Map of Poland showing the location of Ro- agate mortar and pestle. Two samples gowo (geographic coordinates: 53° 4′ 41″ N, were run in duplicate. The mineral com- 18° 44′ 18″ E) position of all 29 apatite samples was ex- amined using Fourier transform infrared spectroscopy (FTIR) to check for diagen- esis, a technique described in greater de- tail by Wright and Schwarcz (1996). The two FTIR measurements used to evalu- ate preservation of carbonate are crystal- linity index (CI), and carbonate content as estimated by a carbonate-phosphate Fig. 2. Burial 59 from Rogowo, referred to as a “princess” on account of her rich jewelry. Pho- ratio (C/P). We also consider peaks or tographed by Adam Szwarc shoulders at wavelength 1096 cm–1, in- 6 Laurie J. Reitsema and Tomasz Kozłowski dicative of recrystallization to franco- versity. Collagen samples were analyzed lite, as a window into post-mortem al- on a Costech Elemental Analyzer cou- teration of bone (Wright and Schwarcz pled to a Finnigan Delta IV Plus stable 1996). CI and C/P values far outside the isotope ratio mass spectrometer under range of modern bone suggest either continuous flow using a CONFLO III in- that contaminating substances are pres- terface. Stable carbon and nitrogen meas- ent in the sample, or that the sample has urements were made where the average recrystallized during deposition or sam- standard deviation of repeated meas- ple preparation, both of which may shift urements of the USGS24 and IAEA-N1 the isotope signatures in carbonate away standards were between 0.04‰ and δ13 from biogenic values (Berna et al. 2004; 0.07‰ for Ccoll and between 0.07‰ δ15 δ13 Nagy et al. 2008; Szostek 2009; Szostek and 0.18‰ for N. We report Ccoll et al. 2011; Yoder and Bartelink 2010). values to the nearest 0.1‰. We also examine the stable oxygen iso- For carbonate, a 1.0–1.2 mg sub- δ18 δ13 tope ratios ( O) of all apatite samples sample was analyzed for Cap relative as a complementary proxy indicator of to the Vienna Peedee Belemnite Lime- diagenetic alteration, after Wright and stone standard using an automated Car- Schwarcz (1996) and Garvie-Lok (2001). bonate Kiel device coupled to the mass The rationale behind using δ18O values spectrometer. Samples were acidified as indicators of diagenesis rather than as under vacuum with 100% ortho-phos- δ18 biogenic paleodiet indicators is that O phoric acid, the resulting CO2 cryogeni- values in bone apatite are more affect- cally purified, and delivered to the mass ed than δ13C values by fractionation that spectrometer. Three samples were run occurs during changes between different in duplicate and replicability was within species of carbonate in bone during sam- 0.20‰ for δ13C. The standard deviation ple treatment or diagenesis (Wright and of repeated measurements of limestone δ13 Schwarcz 1996). Whereas the ranges internal standard (NBS-19) for Cap of FTIR values alone do not necessarily was ± 0.03‰. Because of high precision indicate diagenesis per se (considerable for carbonate measurements, we report δ13 scatter has been documented in both Cap values to the nearest 0.01‰. modern and “well-preserved” archae- ological bone; see for example Wright Results and Schwarcz [1996] and Garvie-Lok et al. [2004]), correlations between δ18O The preservation quality of collagen and FTIR values cannot parsimonious- samples was assessed using C:N ratios, ly be explained otherwise (Wright and %C and %N following criteria outlined Schwarcz 1996; Garvie-Lok 2001). An- in DeNiro (1985) (Table 1). All collagen imal bones from medieval Poland were samples, including humans and animals, prepared in the same manner as human exhibit C:N ratios of between 3.2 and bones (for more details on the animal 3.4, %C values of between 19.2% and bones used in this study, see Reitsema 43.6%, and %N values of between 6.9% et al. 2013). and 15.7%. Some of these carbon and ni- All stable isotope analyses were con- trogen concentrations are lower than the ducted in the Stable Isotope Biogeochem- ranges for modern collagen, but these istry Laboratory at the Ohio State Uni- samples nevertheless exhibit acceptable Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 7

C:N ratios, and their stable isotope sig- lower C/P than modern bones, there is natures are not anomalous. no relationship between either FTIR CI The use of FTIR to detect diagen- or C/P and either δ18O or δ13C, which has esis is currently limited because there been used elsewhere as a guideline for is no consensus as to acceptable cutoff identifying diagenetically altered sam- points. We report FTIR results to con- ples (c.f., Garvie-Lok 2001) (Fig. 3). Ten tribute to a growing body of literature of the apatite samples exhibit distinct in this area and compare them to pre- peaks at wavenumber 1096 cm–1 indicat- vious studies, while recognizing that ing francolite in the apatite matrix, mean- FTIR does not provide the same kind ing they are more likely than other sam- of assurance of sample “quality” as do ples to have been diagenetically altered collagen quality indicators. FTIR results (c.f., Katzenberg et al. 2009; Wright and of the present study (Table 1) suggest Schwarcz 1996). Nevertheless, through- that carbonate from Rogowo is reason- out the sample, peaks are not consist- ably well-preserved and likely retains ently associated with anomalous CI, C/P δ13 biogenic Cap signatures. CI values or stable isotope ratios (Fig. 3). In keep- range from 2.7 to 4.2 (mean=3.6±0.4) ing with the recommendations of previ- and C/P ratios range from 0.10 to 0.41 ous researchers (Garvie-Lok 2001; Niels- (mean 0.18±0.07). These values resem- en-Marsh and Hedges 2003a; Wright and ble ranges of acetic-treated archaeologi- Schwarcz 1996), rather than considering cal carbonate estimated to be well-pre- FTIR indices as isolated and finite exclu- served elsewhere (e.g., Katzenberg et al. sion criteria, we consider them alongside 2009). For comparison, modern bones other indicators of diagenetic alteration, treated with acetic acid to mimic archae- including correlations between the the ological sample preparation exhibit CI FTIR data and stable isotope values. In values of between 2.7 and 3.6 (Niel­sen- the following discussion we report statis- Marsh and Hedges 2000b; Wright and tical results of the overall apatite sample Schwarcz 1996). Shemesh (1990) advo- (n=29) as well as a more conservative cates considering 3.8 a maximum value restricted sample (n=13; values itali- for archaeological, acid-treated samples, cized in Table 1) comprising only those above which samples should no longer individuals with CI values of 3.6 or low- be considered well-preserved. This is er, and CI values between 0.15 and 0.70 relatively conservative, as de la Cruz Bal- (Berna et al. 2004; Szostek et al. 2011). tazar (2001) reports a CI range of 2.8–4.0 In addition to medieval Polish fauna, in a well-preserved archaeological sam- Roman era faunal data from Denmark are ple, Katzenberg et al. (2009) consider included for comparison to the human a cut-off of 4.0, and Wright and Schwarcz values from Rogowo in the absence of (1996) report well-preserved values directly asssociated animal bones (Table above 4.0. In terms of C/P, modern bones 2). Jørkov et al. (2010) report that 3 cows treated with acetic acid exhibit values of exhibit a mean δ13C of –21.6±0.7‰ and 0.10–0.24 (Wright and Schwarcz 1996). a mean δ15N of 5.7±0.5‰, and 4 pigs Acetic-treated archaeological C/P values exhibit a mean δ13C of –22.1±0.8‰ and tend to be slightly lower (c.f., Katzenberg a mean δ15N of 8.6±0.9‰. et al. 2009). Although several samples Stable isotope data for humans are nd δ13 from 2 c. Rogowo exhibit higher CI and presented in Table 1. Rogowo ­ Ccoll 8 Laurie J. Reitsema and Tomasz Kozłowski

Fig. 3. Indicators of bone mineral diagenesis considered in the present study include (1) the crystallinity index (FTIR CI), (2) carbonate content in bone, estimated through Fourier transform infrared spectros- copy (FTIR C/P), (3) the presence of a peak at FTIR wavelength 1096 cm–1, and (4) any relationship between FTIR indices and stable carbon (δ13C) or oxygen isotope (δ18O) values from apatite. FTIR CI and C/P are correlated, as expected for bone apatite. Although some absolute FTIR indices are ques- tionable, the fact that no correlations exist among FTIR and stable isotope values suggests the samples may be reasonably well-preserved values range widely from –19.5‰ to values (R2 =0.1463), which could occur –16.4‰ (mean=–17.9±0.7‰) and when 13C-enriched fish (marine or - ana δ15N values range from 8.6‰ to 10.9‰ dromous) are a major source of protein in δ13 (mean=9.7±0.5‰) (Fig. 4). There is the diet (e.g., Ambrose et al. 1997). Cap δ13 δ15 no correlation between Ccoll and N values range from –13.09‰ to –9.69‰ Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 9 - 1 m–

––– Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak Smooth Smooth Smooth Smooth Smooth Smooth Smooth Smooth Smooth Smooth Smooth Smooth Shoulder Shoulder Shoulder Shoulder Shoulder Shoulder Shoulder 1096c – C/P 0.16 0.13 0.13 0.15 0.15 0.14 0.15 0.15 0.16 0.14 0.17 0.14 0.10 0.15 0.14 0.14 0.20 0.41 0.19 0.18 0.33 0.29 0.18 0.19 0.24 0.20 0.15 0.24 0.26 0.18 ±0.07 – CI 3.8 4.1 4.2 3.8 3.9 3.9 4.1 3.9 3.8 4.2 4.0 4.0 3.8 3.8 3.8 4.2 3.4 3.4 3.1 2.7 3.4 3.5 3.3 2.9 3.5 3.0 2.9 3.1 3.6 3.6 ± 0.4 ±0.14 ±0.01 (‰) – ap –9.69 C –11.46 –11.01 –12.49 –11.73 –11.27 –11.98 –11.94 –11.96 –10.40 –12.18 –11.39 –11.32 –12.14 –12.11 –11.51 –12.06 –11.61 –11.32 –10.74 –13.07 –13.09 –12.31 –11.20 –11.55 –10.95 13 δ –11.26 –10.46 –12.17±0.03 -11.60 ± 0.76 (‰) coll –17.5 –18.2 –16.9 –18.5 –18.0 –17.7 –16.9 –16.4 –17.2 –17.0 –17.9 –19.1 –18.2 –19.0 –17.9 –17.7 –18.9 –17.4 –17.4 –19.5 –18.3 –17.3 –18.8 –18.5 –16.8 –17.5 –18.1 C 13 –17.9±0.7 δ –18.2±0.03 –17.7±0.15 –17.9±0.02 (‰) 9.7 9.3 9.8 8.6 9.9 9.4 9.0 9.3 9.5 9.7 9.9 9.7 9.3 9.9 9.9 8.9 8.9 9.2 9.4 9.6 Air 10.7 10.0 10.1 10.1 10.3 10.9 10.4 N 9.7±0.5 15 9.7±0.01 9.9±0.01 9.3±0.04 δ 3.2 3.2 3.3 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.3 3.2 3.3 3.2 3.2 3.2 3.2 3.2 3.3 3.2 3.4 3.3 3.2 3.2 3.3 3.2 3.2 3.2 3.2 C:N 3.2±0.05 %C 39.9 40.1 23.3 26.4 26.3 27.9 28.1 35.0 37.2 37.3 38.7 39.3 22.0 22.1 26.6 32.8 34.3 33.8 35.6 35.6 35.1 35.5 36.4 35.8 36.1 38.9 39.2 19.2 36.9 38.0 33.1±6.1 8.5 9.5 9.6 7.9 8.0 9.6 6.9 %N 14.4 14.5 10.1 10.3 12.7 13.1 13.7 13.8 14.2 11.9 12.0 12.3 12.3 12.5 12.8 12.8 12.9 13.1 13.2 14.0 14.2 13.3 13.5 11.9±2.2 sd: ± Age 22–35 22–35 35–55 35–55 22–35 22–55 22–35 22–35 22–35 35–55 22–35 22–35 35–55 20–35 22–35 20–35 22–35 35–55 35–55 20–35 35–55 20–35 22–35 35–55 22–35 22–35 22–35 22–35 35–>60 >55–60 Mean exhibit FTIR indices these are given special consideration in the text, and excluded from some analyses and excluded FTIR indices these are given special consideration in the text, exhibit F F F F F F F F F F F F F F F M M M M M M M M M M M M M M M italics Sex ing in 56 45 38 23 39 43 48 ID 29* 51* 59* 25* 705 100 708 752 104 680 774 615 694 428 699 339 440 676 743 767 438 780* 688N Table 1. Collagen quality indicators, stable isotope values, and Fourier transform infrared spectroscopy (FTIR) results from all humans. Values appear Values humans. all from results (FTIR) spectroscopy infrared transform Fourier and values, isotope stable indicators, quality Collagen 1. Table 10 Laurie J. Reitsema and Tomasz Kozłowski

Table 2. Comparative faunal baseline δ15N δ13C δ13C ID Species %N %C C:N Air coll ap Reference (‰) (‰) (‰) 816-02 Cow Bos primigenius 14.6 40.4 3.2 7.8 –20.7 –11.41 Present study taurus 834-02 Cow Bos primigenius 13.4 38.0 3.3 7.9 –21.0 – Present study taurus 839-02 Cow Bos primigenius 13.1 36.7 3.3 7.6 –21.5 – Present study taurus 200 Cow Bos primigenius 15.7 43.6 3.2 6.4 –21.0 – Present study taurus 810-02 Cow Bos primigenius 14.4 40.2 3.2 6.4 –21.3 –13.80 Present study taurus 100 Pig Sus scrofa 16.1 44.5 3.2 7.9 –21.6 –14.89 Present study domesticus 119-05 Pig Sus scrofa 12.9 35.6 3.2 5.9 –21.6 –14.71 Present study domesticus 816-02 Pig Sus scrofa 13.2 37.6 3.3 6.8 –19.8 –12.66 Present study domesticus

Pig Sus scrofa 8.6±0.9 –22.1±0.8 Jørkov et al. domesticus (n=4) (2010) Cow Bos primigenius 5.7±0.5 –21.6±0.7 Jørkov et al. taurus (n=7) (2010)

(mean=–11.60±0.76‰). When only of 3.6 or lower and C/P values of be- apatite samples with FTIR CI values tween 0.15 and 0.70 are considered

δ13 δ15 Fig. 4. Stable carbon ( Ccoll) and nitrogen ( N) isotope data from collagen from Rogowo, including 15 males, 5 females buried with jewelry, and 10 females buried without jewelry Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 11

δ13 δ13 Fig. 5. Stable carbon isotope data from carbonate ( Cap) and collagen ( Ccoll) from Rogowo (n=29)

δ13 (n=13), the mean Cap value is similar er, albeit not significantly so (all 15- fe δ13 δ13 ≤ (–11.64±0.93). Ccoll and Cap values males, p=0.157; 7 females with CI 3.6, are modestly positively correlated at Ro- p=0.121). The richest grave at Rogowo gowo (R2=0.3439) (Fig. 5). When the is skeleton R–59-F, a woman buried with δ13 restricted Cap sample of just 13 indi- substantial quantities of jewelry (Fig. 2), δ13 δ13 viduals is used, Ccoll and Cap values which we consider in more detail on ac- remain modestly positively correlated count of her unusual mortuary context. 2 Δ13 (R =0.4482). C(ap-coll) values range This individual exhibits relatively high δ13 from 5.3‰ to 8.1‰ (mean=6.3±0.7‰). Ccoll value (–16.8‰) compared to The δ13C values of men and wom- mean values, but the value is similar to en are significantly different: both the several other females (R-705-F, R-56-F, δ13 δ13 ­ Ccoll and Cap values of females are R-688N-F, R-104-F). Furthermore, her δ13 Δ13 higher (Kruskal-Wallis Test, p=0.014 Cap (–10.95‰), C(ap-coll) (5.9‰), and and p=0.015, respectively). When only δ15N values (9.3‰) are unremarkable. apatite samples with FTIR CI values of 3.6 and are considered (n=13, 7 females Discussion δ13 and 6 males) this Cap difference re- mains (Kruskal-Wallis Test, p=0.032). Women exhibit lower mean δ15N than Diet Reconstruction m­en (Kruskal-Wallis Test, p=0.078). Five female skeletons buried with jew- δ15N values suggest a predominantly ter- δ13 δ13 δ13 elry were analyzed. The Ccoll, Cap and restrial diet at Rogowo. C values indi- δ15 N values of women with and without cate a mixed diet of C3 plants, C4 plants, δ13 jewelry are similar, although Cap of and fish. If a diet-collagen spacing rela- women without jewelry is generally low- tionship of +5‰ is assumed, the pro- 12 Laurie J. Reitsema and Tomasz Kozłowski tein diet of individuals at Rogowo can be anadromous fish consumption. δ13C and inferred to have a δ13C value of –21.4‰ δ15N values from both marine and ana- to –24.5‰. In agreement with this, if dromous fish bones tend to be higher a +12‰ diet-carbonate fractionation than those of terrestrial animals, and factor is applied (c.f., Garvie-Lok 2001: humans eating such fish exhibit a -cor 125–128; Prowse et al. 2005), the δ13C relation between the two isotopes (e.g., values of bulk diets at Rogowo ranged Ambrose et al. 1997). However, a grow- between –21.7‰ and –25.0‰. The ex- ing number of studies reveal that stable pected isotopic “space” of a C3 diet is at isotope ratios of fish bones are highly the lower end of this range, –23.0‰ to variable (c.f., Dürrwächter et al. 2006;

–25.0‰ (Garvie-Lok 2001). Thus, C3 Fuller et al. 2012; Katzenberg et al. 2009; foods appear to have dominated diet at Müldner and Richards 2005; Redfern et Rogowo, but foods with a higher δ13C al. 2010; Reitsema et al. 2013). Conse- signature also contributed to both the quently, a straightforward correlation energy and protein sources of the diet between δ13C and δ15N among fish-eating (namely, millet and marine and/or ana- humans may not always exist. dromous fish). To further explore the issue of fish

A C4 signal could have contributed in diet, we consider Rogowo in relation to diet in two ways: through direct con- to other European samples. In Figure 6, sumption of millet by humans, or indi- collagen data from Rogowo (mean±1σ) rectly through consumption of animals are shown alongside 9 previously report- foddered on millet. Sampling local ani- ed European historical populations. Four mals is the most straightforward way to of these are reported to have consumed resolve this, but it is also possible to ex- a predominantly or exclusively terrestrial th plore whether C4-foddered animals con- diet with only C3 foods: a 4 c. AD Late tributed to human diet by checking for Roman sample from England (Richards a correlation between δ13C and δ15N val- et al. 1998), a 5th–2nd c. BC Greek coloni- ues. A closer look at Figure 4 shows that al sample from Bulgaria (Keenleyside et individuals with δ13C values of greater al. 2006), a 2nd c. sample from England than –18.0‰ exhibited the entire range (Chenery et al. 2010), and a 1st–4th c. AD of δ15N values. The fact that δ13C and δ15N Roman Iron Age sample from Denmark values are not positively correlated at Ro- (Jørkov et al. 2010). Three comparative gowo (R2 =0.1463) supports an interpre- samples consumed predominantly or tation of direct millet consumption, rath- exclusively terrestrial diets that also in- er than C4-foddered animal consumption. cluded C4 foods: several Hallstatt groups δ13 The Cca values from Rogowo suggest from Austria (Le Huray and Schutkowski millet contributed between 0% and 35% 2005), a 9th–5th c. Iron Age sample from to overall diet, depending on the individu- Slovenia (Murray and Schoeninger al. This corroborates archaeobotanical ev- 1988) and an 11th–12th c. AD early me- idence for a broad-spectrum agricultural dieval population from Poland (Reitse- strategy including millet, barley, rye, oats ma et al. 2010). Finally, two comparative and wheat (Wasylikowa et al. 1991). samples reportedly consumed marine δ13 The lack of a correlation between C foods in addition to C3 terrestrial foods: and δ15N values would seem to caution 2nd–5th c. skeletons from Roman Tunisia against an interpretation of marine or (­Keenleyside et al. 2009) and a Roman Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 13 era sample from Sweden (Eriksson et al. general exhibit lower stable isotope val- 2008). ues, humans’ δ15N values at Rogowo are δ15N values place Rogowo near other 4‰ higher than cows and 2‰ higher populations with terrestrial diets. Al- than pigs. Accepting either the Polish or though the δ15N values at Rogowo are the Danish faunal baseline would still in- not the highest among the terrestrial dicate a predominantly terrestrial diet, if diets displayed in Figure 6, cautioning a diet-tissue spacing of 3–5‰ is accepted against an interpretation of a high-fish for δ15N (Drucker and Bocherens 2004). δ13 diet, they are high in comparison to the In terms of C values in Figure 6, lower cluster of populations. For exam- Rogowo overlaps with Hallstatt popu- δ15 ple, the mean N value from another lations in Austria that consumed a C4 th th Polish sample dating to the 11 –12 c. plant, millet. However, C3 plants still δ13 is 9.2±0.6‰ (approximately 3‰ high- predominated in the diet: Rogowo Ccoll er than its faunal baseline) (Reitsema values are considerably lower than those et al. 2010). δ15N values at Rogowo of a Slovenian population that consumed (9.7±0.5‰) are higher than the Giecz large amounts of millet (or millet-fod- sample. Interestingly, humans’ δ15N val- dered animals). Millet has also been de- ues at Rogowo are still just 2.5‰ high- tected isotopically in human remains in er than domestic animal values reported early medieval Poland (Reitsema 2012; by Reitsema et al. (2013) from nearby Reitsema et al. 2010), Bronze Age Ita- Kałdus, a medieval settlement approx- ly (Tafuri et al. 2009), Iron Age Austria imately 1000 years younger than Ro- and Slovenia (Le Huray and Schutkowski gowo. A diet-tissue difference of 2.5‰ 2005; Murray and Schoeninger 1988), is a reasonable expectation for consum- and possibly as early as the Neolithic in ers of terrestrial animal protein. As pre- Poland (Kozłowski et al. 2013). Unlike in viously discussed, the medieval animals Eastern and Southern Europe, millet is from Kałdus exhibit unusually high δ15N not widespread in Western Europe. With values likely due to land management its occurrence following cultural distri- practices (e.g., manuring, ploughing, bution patterns rather than geography fire-clearing, etc.). At some point in an- or climate, millet appears to characterize tiquity, the local baseline δ15N in this re- diets of Slavic populations. In her book gion of Poland may have been raised by documenting the history of Polish cui- human land management strategies such sine, Maria Dembińska posits that “… as burning or fertilizing fields (Bogaard the Polish preference for millet was cul- et al. 2007; Commisso and Nelson 2008; tural and that millet must have arrived Grogan et al. 2000), and it is possible with the earliest Slavic migrants” (104). this occurred as early as the Roman era. That millet might be used as a proxy for A grain of wheat sampled from a nearby Slavic migrations is an interesting possi- modern field yielded a highδ 15N value of bility, given the controversy surrounding 7.2‰, which could be due to the influ- the relationship between the Wielbark ence of either modern or past land man- culture and the issue of Slavic ethno- agement strategies on soil nitrogen (c.f., genesis (c.f., Dąbrowski 2006). Millet Commisso and Nelson 2008). Compared consumption among Wielbark and early to the Roman era animals from Denmark medieval Polish populations represents reported by Jørkov et al. (2010) which in a potentially significant continuity - be 14 Laurie J. Reitsema and Tomasz Kozłowski

δ13 δ15 Fig. 6. Collagen stable isotope data (carbon: Ccoll; nitrogen: N) from Rogowo (n=30) compared to other European populations tween the Roman era and the medieval an populations after shifting the regres- δ13 period in Poland. sion lines by +1.5‰ for both Ccoll and δ13 δ13 A model for using both C from car- Cap to account for the Suess effect, by bonate and collagen was developed by which the modern-day atmospheric δ13C Kellner and Schoeninger (2007) to pro- value is lower than in the past due to re- vide a clearer idea of protein and ener- cent fossil fuel emissions (Marino and gy in diet. The model consists of three McElroy 1991). All 29 apatite samples regression lines that were derived from are included but the 16 samples whose controlled feeding experiments with an- FTIR CI values exceed 3.6 and C/P val- imals. Each line represents a particular ues are below 0.15 are demarcated with diet: one based on C3 protein, one based an “X”. The importance of millet in diet on marine protein, and one based on C4 is supported by Rogowo’s location at the protein. The end points of each line rep- middle of the C3 protein line, rather than resent either 100% C3 energy (bottom) or towards the base where “C3-only” popu-

100% C4 energy (top). Using this model, lations would plot. The fact that Rogowo it is possible to tease apart sources of sta- plots to the right of the C3 protein line ble isotope variation from small amounts indicates that dietary protein came from of fish or millet. In Figure 7, both types primarily C3 terrestrial fauna, but with of δ13C data from Rogowo are displayed supplemental input from marine or an- in comparison to several other Europe- adromous fish. Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 15

Few comparative populations are tein line. Also to the left of the line are available for use in Figure 7 because col- populations from inland Greece where lagen and carbonate are not yet common- marine foods were negligible. Rogowo ly analyzed together in stable isotope appears very similar to Roman era Tuni- studies of Europe. Pictured are: six pop- sia, where fish made significant contri- ulations from Neolithic Greece including butions to diet (Keenleyside et al. 2009). three from coastal areas and three from Interestingly, the δ15N values from Ro- further inland (Papathanasiou 2003), gowo are considerably lower than those a medieval Polish population (Reitsema from the Tunisian sample (9.7±0.5‰ vs. et al. 2010), Roman populations from 12.8±1.3‰). We conclude that anadro- Tunisia (Keenleyside et al. 2009) and Ita- mous or marine fish were consumed in ly (Prowse et al. 2005), and humans from supplemental amounts at Rogowo, and Viking era and medieval Sweden (Kosiba­ that they probably had relatively low δ15N et al. 2007). Populations consuming fish values. A “fish signal” thus emerges only plot to the right of the C3 protein line in when using the model from Kellner and this figure with one exception: the Ro- Schoeninger (2007), and is not apparent man Italian sample from Prowse et al. when examining collagen data alone. (2005) reportedly consumed some ma- The nearby Vistula River, along with its rine foods, but plots left of the C3 pro- tributaries and oxbow lakes, would have

δ13 δ13 Fig. 7. Stable carbon isotope data from carbonate ( Cap) and collagen ( Ccoll) from Rogowo (n=29) compared to other European populations. Regression lines are from Kellner and Schoeninger (2007). Because the regression lines were developed using modern animal bones, which will exhibit systemat- δ13 ically lower C values due to differences between modern and pre-industrial atmospheric CO2 values stemming from fossil fuel emissions (Marino and McElroy 1991), regression lines are shifted to be 1.5‰ higher to make them comparable with the archaeological human samples. Sixteen samples ex- hibiting crystallinity indices (FTIR CI) in excess of 3.6, and carbonate content (FTIR C/P) below 0.15 are demarcated with a white “X” through the data symbol 16 Laurie J. Reitsema and Tomasz Kozłowski provided the population at Rogowo with ties among the population. Women and freshwater (carp, bream, roach) and an- men may have performed different sub- adromous (sturgeon) fish. Oxbow lakes sistence tasks throughout the day, simply composed of shallow, still water were exposing women and men to different also present in the region which may foods. Increased consumption of plants contain fish with very low δ15N values in general would augment a “millet sig- and relatively high δ13C values (Hecky nal” in bones of women at Rogowo, be- and Hesslein 1995). Three anadromous cause it would no longer be so strongly fish (sturgeon) from another nearby me- masked the available C3 protein from ter- dieval settlement, Kałdus, show δ13C val- restrial animals. ues of –15.6‰ to –17.1‰ and of 9.9‰ It is also possible that the δ13C dif- to 11.3‰ (Reitsema et al. 2013). Con- ferences between men and women are sumption of these fish would be fairly non-dietary. Plants from warm climates consistent with the human stable isotope have been shown to exhibit higher δ13C data we have reported. However, their values than the same plants in colder contribution to diet was probably sup- climates due to temperature-dependent plemental. Other sotopic paleodietary fractionation of CO2 (van Klinken et al. research has demonstrated that Europe- 2000). Richards et al. (1998) identified an populations in the first centuries AD possible immigrants at the Poundbury were not exploiting aquatic resources site in England based on the individu- though they presumably had the means als’ elevated δ13C values. Perhaps some (e.g., Chenery et al. 2010; Jørkov et al. women with higher δ13C values migrated 2010; Richards et al. 1998). Like these to Rogowo from a warmer climate. Us- other populations, the subsistence econ- ing morphometrics, Gładykowska-Rze- omy at Rogowo was based on animal czycka et al. (1997) report that females husbandry and agriculture. buried in another Wielbark cemetery in Pruszcz Gdański were an “autochthonic Intrapopulation Variation population of ” whereas males had “traces characteristic for Goths-German Sex-based differences in stable isotope tribes that [moved into] the East Pomer- ratios at Rogowo suggest that diets of ania region” (Gladykowska-Rzeczycka et women included more millet, and pos- al. 1997:92), suggesting sex-biased resi- sibly less animal protein (meat or dairy) dency patterns. However, contra the evi- than diets of men. Sex-based social dif- dence from Pruszcz Gdański, if migration ferentiation is a characteristic of many underlies the stable isotope differences tribal societies and at Rogowo, appears in the present study it would suggest to have translated into differential food women were the immigrants at Rogowo. consumption. Sex-based differences in Analyses of strontium and oxygen iso- diet from this time period are reported tope values of tooth enamel and bone by other stable isotope and trace element from Rogowo are intended to further ex- studies from this time period in Europe plore questions of mobility. (e.g., Fuller et al. 2006; Redfern et al. Stable isotope differences were not 2010; Richards et al. 1998; Smrćka et al. associated with differences in grave 2000). However, such differences do not goods (jewelry) in women’s burials (Fig. necessarily implicate sex-based inequali- 4). This accords with an interpretation of Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 17

Wielbark populations as relatively egal- and terrestrial animals are not always itarian, with biological variables (e.g., discrete. The stable isotope ratios of fish sex) playing a more influential role in and terrestrial foods frequently overlap, structuring diet than cultural variables rendering fish isotopically invisible in (e.g., status). human diet. Even more complications arise when past land management strat- Conclusion egies such as manuring, ploughing and burning fields raise δ15N values at the Diet at Rogowo was based on terrestrial base of the human food web, or when foods (meat, milk, and plants), but with human consumption of legumes such supplemental input from fish (likely low- as peas lowers δ15N values. Because of δ15N anadromous or littoral fish), indi- these obstacles it is helpful to consider cating an overall broad-spectrum sub- relationships between different stable sistence economy. Millet was consumed isotope ratios (δ15N, and δ13C from both at Rogowo, and in greater amounts than collagen and apatite) and other patterns during the medieval period to follow in the data, rather than simply compar- (c.f., Fig. 6; Reitsema et al. 2010; Reit- ing stable isotope ratios obtained from sema 2012). Stable isotope differences human bones to expected ranges. Con- between the sexes indicate that men and siderations such as these have permit- women may have consumed different ted the identification of millet and ana- foods, although there is a possibility that dromous or marine fish consumption at women migrated to Rogowo from an iso- Rogowo. Stable carbon and nitrogen iso- topically different region of Europe. No tope data accord with paleobotanical and significant differences are noted in δ13C archaeozoological evidence for a mixed or δ15N of burials with and without jew- subsistence economy of agriculture and elry, suggesting low status-based social animal husbandry, and contribute new differentiation among the Wielbark cul- information regarding the roles of biol- ture, at least in terms of daily food ac- ogy and culture in structuring diet and cess. possibly social structure of the relatively A significant obstacle to paleodiet unknown Wielbark culture. studies in Europe using stable isotope evidence is that most foods are isotop- Author contribution ically indistinguishable. Consumption of the same foods is not necessarily im- LJR prepared and analysed bone samples plicated when individuals exhibit sim- isotopically, obtained grant sponsorship, ilar stable isotope signatures. In some and wrote the manuscript; TK contribut- Eastern and Southern European regions, ed bone material for the study, provided millet creates isotopic heterogeneity that archaeological and contextual informa- may highlight differential access within tion about the site, and edited drafts of the population. This is the case at Ro- the manuscript. gowo, but not in Western Europe where agriculture was entirely based on isotop- Conflict of interests ically homogeneous C3 plants. A second problem with diet reconstruction in Eu- The authors declare that there is no con- rope is that stable isotope ratios of fish flict of interests. 18 Laurie J. Reitsema and Tomasz Kozłowski

Acknowledgments bonate. In: Prehistoric human bone: ar- chaeology at the molecular level. Lambert We thank Andréa Grottoli and Yohei JB and Grupe G, editors. Berlin: Spring- Matsui at The Ohio State University Sta- er-Verlag. 1–37. Barford PM. 2001. The early : culture ble Isotope Biogeochemistry Laboratory, and society in early medieval Eastern Eu- Rachel Kopec and Stephen Schwartz at rope. Ithaca: Cornell University Press. The Ohio State University Department Berna F, Matthews A, Weiner S. 2004. Solu- of Food Sciences and Technology, Daniel bilities of bone mineral from archaeolog- Makowiecki at the Nicolaus Copernicus ical sites: The recrystallization window. J University, and Douglas E. Crews at The Archaeol Sci 31:867–82. Ohio State University Human Biology Bocherens H, Drucker D. 2003. Trophic Level Laboratory. Isotopic Enrichment of Carbon and Nitro- gen in Bone Collagen: Case Studies from Recent and Ancient Terrestrial Ecosys- Corresponding author tems. Int J Osteoarchaeol 13:46–53. Bogaard A, Heaton THE, Poulton P, Merbach Laurie J. Reitsema, Department of An- I. 2007. The impact of manuring on nitro- thropology, University of Georgia, 250 gen isotope ratios in cereals: Archaeologi- Baldwin Hall, Jackson St., Athens Geor- cal implications for reconstruction of diet gia 30602, USA and crop management practices. J Archae- e-mail address: [email protected] ol Sci 34:335–43. Britton K, Müldner G, Bell M. 2008. Stable isotope evidence for salt-marsh grazing References in the Bronze Age Severn Estuary, UK: implications for palaeodietary analysis at Ambrose SH. 1990. Preparation and charac- coastal sites. J Archaeol Sci 35:2111–18. terization of bone and tooth collagen for Buko A. 2008. The Archaeology of Early Me- isotopic analysis. J Archaeol Sci 17:431– dieval Poland: Discoveries - Hypotheses - 51. Interpretations. Boston: Brill. Ambrose SH. 1991. Effects of diet, climate Chenery C, Müldner G, Evans J, Eckardt H, and physiology on nitrogen isotope abun- Lewis M. 2010. Strontium and stable dances in terrestrial foodwebs. J Archaeol isotope evidence for diet and mobility in Sci 18:293–317. Roman Gloucester, UK. J Archaeol Sci Ambrose SH. 2000. Controlled diet and cli- 37:150–63. mate experiments on nitrogen isotope ra- Chisholm BS, Nelson DE, Schwarcz HP. 1982. tios of rats. In: Biogeochemical approach- Stable-carbon isotope ratio as a measure es to paleodietary analysis. Ambrose SH of marine versus terrestrial protein in an- and Katzenberg AM, editors. New York: cient diets. Science 216:1131–32. Kluwer Academic/Plenum Publishers. Chudziak W. 2000. Archeologia na bydgosko- 243–59. -toruńskim odcinku autostrady A1. Z ot- Ambrose SH, Butler BM, Hanson DB, Hunt- chłani wieków. Archeologia na dużych in- er-Anderson RL, Krueger HW. 1997. Sta- westycjach w Polsce 55:28–32. ble isotopic analysis of human diet in the Commisso RG, Nelson DE. 2008. Correlation Marianas Archipelago, Western Pacific. between modern plant d15N values and Am J Phys Anthropol 104:343–61. activity areas of Medieval Norse farms. J Ambrose SH, Norr L. 1993. Experimental ev- Archaeol Sci 35:492–504. idence for the relationship of the carbon isotope ratios of whole diet and dietary protein to those of bone collagen and car- Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 19

Czekanowski J. 1955. Goci a Lechici i dowody chaeological sites (1st and 2nd millenni- antropologiczne. Przgl Antropol 21:857– um AD). J Anal At Spectrom 27:807–20. 91. Garvie-Lok S. 2001. Loaves and fishes: a sta- Dąbrowski R. 2006. Origin of the Slavs: an ble isotope reconstruction of diet in medi- anthropological perspective. Archaeologia eval Greece. PhD Thesis, Department of Polona 44:333–38. Anthropology, University of Calgary. 554 DeNiro MJ, Epstein S. 1978. Influence of di- p. et on the distribution of carbon isotopes Garvie-Lok SJ, Varney TL, Katzenberg MA. in animals. Geochim Cosmochim Ac 2004. Preparation of bone carbonate 42:495–506. for stable isotope analysis: the effects of DeNiro MJ, Epstein S. 1981. Influence of diet treatment time and acid concentration J on the distribution of nitrogen isotopes Archaeol Sci 31:763–76. in animals. Geochim Cosmochim Ac Gieysztor A, Kiniewicz S, Rostworowski E, 45:341–51. Tazbir J, Wereszycki H. 1979. History of Drucker D, Bocherens H. 2004. Carbon and Poland. Warsaw: PWN – Polish Scientific nitrogen isotopes as tracers of change in Publishers. diet breadth during Middle and Upper Gładykowska-Rzeczycka JJ, Smrčka V, Jambor Paleolithic in Europe. Int J Osteoarchaeol J. 1997. Reconstruction of diet of popula- 14:162–77. tion from Roman period (1st–4th c. A.D.) Drucker DG, Bocherens H, Billiou D. 2003. on Pruszcz Gdański cemetery material. Evidence for shifting environmental con- Przegl Antropol 60:89–95. ditions in Southwestern France from Godłowski K, Kozłowski J. 1985. Historia sta- 33,000 to 15,000 years ago derived from rożytna ziem polskich. Warsaw: Państwo- carbon–13 and nitrogen–15 natural abun- we Wydawnictwo Naukowe. dances in collagen of large herbivores. Grogan P, Bruns TD, Chapin FS. 2000. Fire Earth Planet Sci Lett 216:163–73. effects on ecosystem nitrogen cycling in Dufour E, Herve B, Mariotti A. 1999. Paleo- a Californian bishop pine forest. Oecolo- dietary implications of isotopic variability gia 122:537–44. in Eurasian lacustrine fish. J Archaeol Sci Grupe G, Heinrich D, Peters J. 1999. A brack- 26:617–27. ish water aquatic foodweb: trophic levels Dürrwächter C, Craig OE, Collins MJ, Burger and salinity gradients in the Schlei fjord, J, Alt KW. 2006. Beyond the grave: varia- Northern Germany, in Viking and medie- bility in Neolithic diets in Southern Ger- val times. J Archaeol Sci 36:2125–44. many? J Archaeol Sci 33:39–48. Heather P. 1996. The Goths. Cambridge: Eriksson G, Linderholm A, Fornander E, Kan- Blackwell Publishers Inc. strup M, Schoultz P, Olofsson H, Lidén K. Heaton T. 1999. Spatial, Species, and Tem- 2008. Same island, different diet: cultural poral Variations in the 13C/12C Ratios evolution of food practice on Öland, Swe- of C3 Plants: Implications for Palaeodiet den, from the Mesolithic to the Roman Studies. J Archaeol Sci 26:637–49. Period. J Anthropol Archaeol 27:520–43. Hecky RE, Hesslein RH. 1995. Contributions Fuller BT, Fuller JL, Sage NE, Harris DA, of benthic algae to lake food webs as re- O’Connell TC, Hedges REM. 2005. Nitro- vealed by stable isotope analysis. J N Am gen balance and δ15N: why you’re not what Benthol Soc 14:631–53. you eat during nutritional stress. Rapid Hobson KA, Alisauskas RT, Clark RG. 1993. Commun Mass Spectrom 19:2497–506. Stable-nitrogen isotope enrichment in avi- Fuller BT, Muldner G, Van Neer W, Ervynck A, an tissues due to fasting and nutritional Richards MP. 2012. Carbon and nitrogen stress: implications for isotopic analysis stable isotope ratio analysis of freshwater, of diet. The Condor 95:388–94. brackish and marine fish from Belgian ar- 20 Laurie J. Reitsema and Tomasz Kozłowski

Jørkov MLS, Jørgensen L, Lynnerup N. 2010. Wydawnictwo Uniwersytetu Marii Curie- Uniform diet in a diverse society. Reveal- -Skłodowskiej. ing new dietary evidence of the Danish Kozłowski T, Stepańczak B, Reitsema LJ, Roman Iron Age based on stable isotope Osipowicz G, Szostek K, Płoszaj T, Jędry- analysis. Am J Phys Anthropol 143:523– chowska-Dańska K, Pawlyta J, Paluszkie- 33. wicz C, Witas HW. In Press. Osteologi- Juras A. 2012. Etnogeneza Słowian w świe- cal, chemical and genetic analyses of the tle badań kopalnego DNA. Ph.D. Thesis. human skeletal remains from a Neolithic Adam Mickiewicz University: Poznań. site representing the Globular Amphora Katzenberg AM. 1993. Age differences and Culture (Kowal, Kuyavia region, Poland). population variation in stable isotope val- Anthropologie. ues from Ontario, Canada. In: Prehistoric Krenz-Niedbała M, Kozłowski T. 2013. Com- human bone – anthropology at the molec- paring the chronological distribution of ular level. Lambert JB and Grupe G, edi- enamel hypoplasia in Rogowo, Poland tors. Berlin: Springer-Verlag. 39–61. (2nd century AD) using two methods of Katzenberg AM. 2008. Stable isotope analy- defect timing estimation. Int J Osteoar- sis: a tool for studying past diet, demog- chaeol 23:410–20. raphy, and life history In: Biological An- Larsen CS. 1997. Bioarchaeology: Interpret- thropology of the Human Skeleton, 2e. ing Behavior from the Human Skeleton. Katzenberg AM and Saunders SR, editors. Cambridge: Cambridge University Press. Hoboken: Wiley-Liss. 413–60. Le Huray JD, Schutkowski H. 2005. Diet and Katzenberg AM, Goriunova OI, Weber AW. status during the La Téne period in Bohe- 2009. Paleodiet reconstruction of Bronze mia: Carbon and nitrogen stable isotope Age Siberians from the mortuary site of analysis of bone collagen from Kutná Ho- Khuzhir-Nuge XIV, Lake Baikal. J Archae- ra-Karlov and Radovesice. J Anthropol Ar- ol Sci 36:663–74. chaeol 24:135–47. Keenleyside A, Schwarcz H, Panayotova K. Makowiecki D. 2006. Archaeozoology’s con- 2006. Stable isotopic evidence of diet in a tribution to the improvement of histori- Greek colonial population from the Black ans’ conceptions of subsistence economy Sea. J Archaeol Sci 33:1205–15. and environment in Early Medieval Po- Keenleyside A, Schwarcz H, Stirling L, Lazreg land – selected problems. In: Beiträge zur NB. 2009. Stable isotopic evidence for diet Archäozoologie und Prähistorische An- in a Roman and Late Roman population thropologie 5. Benecke N, editor. 77–82. from Leptiminus, Tunisia. J Archaeol Sci Marino BD, McElroy MB. 1991. Isotopic com- 36:51–63. position of atmospheric CO2 inferred Kellner CM, Schoeninger MJ. 2007. A simple from carbon in C4 plant cellulose. Nature carbon isotope model for reconstructing 349:127–31. prehistoric human diet. Am J Phys An- Müldner G, Richards MP. 2005. Fast or feast: thropol 133:1112–27. reconstructing diet in later medieval Eng- Kosiba SB, Tykot RH, Carlsson D. 2007. Sta- land by stable isotope analysis. J Archaeol ble isotopes as indicators of change in the Sci 32:39–48. food procurement and food preference of Murray ML, Schoeninger MJ. 1988. Diet, sta- and Early Christian popula- tus, and complex social structure in Iron tions on (Sweden). J Anthropol Age Central Europe: some contributions Archaeol 26:394–411. of the bone chemistry. In: Tribe and polity Kozak-Zychman W. 1996. Charakterystyka in late prehistoric Europe. Gibson DB and antropologiczna ludności Lubelszczyzny Geselowitz MN, editors. New York: Ple- z młodszego okresu rzymskiego. Lublin: num Press. 155–76. Isotopic paleodiet analysis at 2nd c AD Rogowo, PL 21

Nielsen-Marsh CM, Hedges REM. 2000a. Pat- Reitsema LJ. 2012. Stable Isotope Evidence terns of diagenesis in bone I: effects of site for Human Diet Change in Poland. Ph.D. environment. J Archaeol Sci 27:1139–50. Thesis. The Ohio State University. 389 pp. Nielsen-Marsh CM, Hedges REM. 2000b. Reitsema LJ, Crews DE, Polcyn M. 2010. Pre- Patterns of diagenesis in bone II: effects of liminary evidence for medieval Polish diet acetic acid treatment and the removal of from carbon and nitrogen stable isotopes. 2- diagenetic CO3 . J Archaeol Sci 27:1151– J Archaeol Sci 37:1413–23. 59. Reitsema LJ, Kozłowski T, Makowiecki D. Papathanasiou A. 2003. Stable isotope anal- 2013. Human-environment interactions ysis in Neolithic Greece and possible im- in medieval Poland: A perspective from plications on human health. Int J Osteoar- the analysis of faunal stable isotope ratios. chaeol 13:314–24. J Archaeol Sci 40:3636–46. Park R, Epstein S. 1960. Carbon isotope frac- Reitsema LJ, Vercellotti G. 2012. Stable iso- tionation during photosynthesis. Geo- tope evidence for sex- and status-based chim Cosmochim Ac 21:110–20. variations in diet and life history at me- Park R, Epstein S. 1961. Metabolic fractiona- dieval Trino Vercellese, Italy. Am J Phys tion of C13 and C12 in plants. Plant Physiol Anthropol 148:589–600. 36:133–28. Richards MP, Hedges REM, Molleson TI, Piontek J, Iwanek B, Segeda S. 2008. Antro- Vogel JC. 1998. Stable isotope analysis pologia o pochodzeniu Słowian. Poznań: reveals variations in human diet in the Monografie Instytutu Antropologii UAM. Poundbury Camp cemetery site. J Archae- Piontek J, Segeda S, Iwanek B, Kozłowski T. ol Sci 25:1247–52. 2006. Odontological analysis of the Wiel- Schoeninger MJ. 2011. Diet reconstruction bark culture population from Rogowo and ecology using stable isotope ratios. cemetery, Poland. Česká Antropologie In: A Companion to Biological Anthro- 56:102-04. pology. Larsen CS, editor. Chichester: Polet C, Katzenberg MA. 2003. Reconstruc- Wiley-Blackwell. tion of the diet in a mediaeval monastic Schoeninger MJ, DeNiro MJ. 1984. Nitrogen community from the coast of Belgium. J and carbon isotopic composition of bone Archaeol Sci 30:525–33. collagen from marine and terrestrial ani- Privat KL, O’Connell TC, Richards MP. 2002. mals. Geochim Cosmochim Ac 48:625– Stable isotope analysis of human and fau- 39. nal remains from the Anglo-Saxon cem- Schoeninger MJ, Moore K. 1992. Bone stable etery at Berinsfield, Oxfordshire: dietary isotope studies in archaeology. J World and social implications. J Archaeol Sci Prehist 6:247–96. 29:779–90. Shemesh A. 1990. Crystallinity and diagene- Prowse T, Schwarcz H, Saunders SR, Mac- sis of sedimentary apatites. Geochimica et chiarelli R, Bondioli L. 2005. Isotopic Cosmochimica Acta 54:2433–38. evidence for age-related variation in diet Smith BN, Epstein S. 1971. Two categories from Isola Sacra, Italy. Am J Phys An- of 13C/12C ratios for higher plants. Plant thropol 128:2–13. Physiol 47:380–84. Pyrgała J. 1975. The reconstruction of agricul- Szostek K. 2009. Chemical signals and recon- ture and breeding economy in Płock Ma- struction of life strategies from ancient zovia at the decline of Antiquity. Archaeo- human bones and teeth – problems and logia Polona 16:71–106. perspectives. Anthropol Rev 72:3–30. Redfern RC, Hamlin C, Athfield NB. 2010. Szostek K, Stepańczak B, Szczepanek A, Kępa Temporal changes in diet: a stable isotope M, Głąb H, Jarosz P, Włodarczak P, Tunia analysis of late Iron Age and Roman Dor- K, Pawlyta J, Paluszkiewicz Cz, Tylko G. set, Britain. J Archaeol Sci 37:1149–60. 2011. Diagenetic signals from ancient hu- 22 Laurie J. Reitsema and Tomasz Kozłowski

man remains – Bioarchaeological applica- Analysis. Ambrose SH and Katzenberg tions. Mineralogia 42:93–112. AM, editors. New York: Kluwer Academ- Tafuri MA, Craig OE, Canci A. 2009. Stable ic/Plenum Publishing. 39–63. isotope evidence for the consumption of Wasylikowa K, Carciumaru M, Hajnalova E, millet and other plants in Bronze Age Ita- Hartyanyi BP, Pashkevich GA, V. YZ. 1991. ly. Am J Phys Anthropol 139:146–53. East-Central Europe. In: Progress in old Tieszen LL, Fagre T. 1993. Effect of diet world palaeoethnobotany; a retrospective quality and composition on the isotopic view on the occasion of 20 years of the

composition of respiratory CO2, bone col- International Work Group for Palaeoeth- lagen, bioapatite, and soft tissues. In: Mo- nobotany. Van Zeist W, Wasylikowa K and lecular Archaeology of Prehistoric Human Behre KE, editors. Balkema, Rotterdam: Bone. Lambert JB and Grupe G, editors. Taylor & Francis. 207–39. Berlin: Springer. 121–55. Wołgiewicz R. 1981. Kultura oksywska i wiel- van der Merwe NJ, Vogel JC. 1978. 13C Con- barska. In: Prahistoria ziem polskich Póź- tent of human collagen as a measure of ny okres leteński i rzymski. Wrocław, prehistoric diet in late woodland North Warszawa, Kraków, Gdańsk: Ossolineum. America. Nature 276:815–16. 135–90. van Klinken GJ, Richards MP, Hedges REM. Yoder C, Bartelink E. 2010. Effects of differ- 2000. An overview of causes for stable ent sample preparation methods on stable isotopic variations in past European hu- carbon and oxygen isotope values of bone man populations: environmental, eco- apatite: A comparison of two treatment physiological, and cultural effects. In: Bi- protocols. Archaeometry 52:115–30. ogeochemical Approaches to Paleodietary