You are on page 1of 9

Research Article

Received: 29 July 2011 Revised: 3 November 2011 Accepted: 7 November 2011 Published online in Wiley Online Library

Rapid Commun. Mass Spectrom. 2012, 26, 327335 (wileyonlinelibrary.com) DOI: 10.1002/rcm.5323

Studying ancient crop provenance: implications from d13C and d15N values of charred barley in a Middle Bronze Age silo at Ebla (NW Syria)
Girolamo Fiorentino1*, Valentina Caracuta1, Grazia Casiello2, Francesco Longobardi2 and Antonio Sacco2
1 2

Laboratory of Archaeobotany and Palaeoecology, University of Salento, - Cultural Heritage, Via D. Birago 64, Lecce 73100 Italy Department of Chemistry, University of Bari, Via Orabona 4, Bari 70126 Italy

The discovery of a storeroom full of barley and other cereals (L.9512) in the proto-historic site of Ebla has provided a unique opportunity to study the centralized storage system of the early city-state from a different perspective. Epigraphic evidence available within the site reveals a complex system of taxation which included gathering grain tributes from satellite sites and redistributing semi-nished products such as our. In this paper, we intend to explore the possibilities of a combined approach to studying the storage system, based on estimated barley grain volumes and d13C-d15N analyses. This approach is used to distinguish between grain from different harvesting sites and to identify any grain cultivated using special agricultural practices (e.g. manuring or irrigation). The basic assumption for this kind of analysis is that the growth-site conditions, natural or anthropogenic, of harvested cereals are reected in their grain size and d13C-d15N values. Since the remains found in the storeroom were charred, the rst task was to evaluate the effect of carbonization on the d13C-d15N and the size of the grains. Thus, the effect of charring was tested on modern samples of Syrian barley landraces. Once it had been ascertained that fresh grains reduced to charred remains retain their original biometric and isotopic traits, the ancient material was examined. Thirteen groups were identied, each characterized by a specic average volume and specic carbon and nitrogen values. The analysis revealed that what had rst appeared to be a homogeneous concentration of grain was in fact an assemblage of barley harvested from different sites. Copyright 2012 John Wiley & Sons, Ltd.

Stable isotopes of nitrogen and carbon have been widely used in archaeology to investigate the dietary habits of ancient populations,[1,2] past environmental conditions[3] and agricultural practices.[4,5] For the latter, most of the analyses have focused on irrigation,[6] fertilization techniques and the state of soils.[7] Given that these two isotopes have rarely been studied jointly in previous analyses in archaeological issues, by combining isotopic ratio data (d13C and d15N values) and size analysis we intend to characterize growing-site conditions and/or identify agricultural practices (e.g. irrigation or manuring) which can affect isotope values in plants. Biometric analysis helps to determine the origin of plant remains: grain size is affected by edaphic conditions during growth,[8,9] and therefore grains of the same species from the same harvesting site are likely to have the same volume.[1012] Stable isotope values help to rene this analysis by characterizing the environmental features of the harvesting site.[13,14] In addition, they can indicate additional water or nutrient inputs.[15,16]

In this study, we selected 91 caryopses of barley from the archaeobotanical assemblage found in storeroom L.9512. Wheat was the main stored cereal but barley was preferred because it was expected to better reect growth site conditions.[17,18] Since the archaeobotanical assemblage in L.9512 was made up of charred remains, the rst task was to evaluate the effect of carbonization on the isotope signature and the size of the barley grains. Thus, the effect of charring was tested on modern samples of Syrian barley landraces. Once it was ascertained that fresh grains reduced to charred remains more or less retain their characteristic size and d13C-d15N values, we plotted these parameters together in order to characterize the different harvesting sites of the barley in the ancient material and to verify whether further distinctions could be made in terms of agricultural practices.

RESEARCH BACKGROUND
The main chemical components of cereals are primarily governed by the water and soil layer in which they grow and can to some extent be inuenced by human practices such as additional water inputs and manuring.[19] Therefore, if used properly and correctly understood, stable carbon and nitrogen isotopes can provide reliable information on

* Correspondence to: G. Fiorentino, Laboratory of Archaeobotany and Palaeoecology, University of Salento, Cultural Heritage, Via D. Birago 64, Lecce 73100, Italy. E-mail: girolamo.orentino@unisalento.it
Rapid Commun. Mass Spectrom. 2012, 26, 327335

327

Copyright 2012 John Wiley & Sons, Ltd.

G. Fiorentino et al. ancient plant water status[6] and plant growing conditions.[20] The great potential of these parameters has been partially explored, but some clarications are required before deriving any valuable conclusions. Previous studies have demonstrated that d15N can be used as an indicator of the nutritional status of ancient crops, since it reects external N sources and the 15N/14N fractionations which occur during the assimilation, transport and loss of N.[21] The main N input is not easy to determine since the isotope ratio in plants can result from both N xation and atmospheric deposition. However, this should not be an issue in arid lands since it has been demonstrated that more than 98% of the organic nitrogen pool is found in the soil, where it is xed by microrganisms.[22,23] Moreover, in natural N-limited systems, uptake of N has to be very efcient, resulting in virtually all available inorganic N being taken up, and, therefore, in no or negligible loss of N.[24,25] Another factor which inuences d15N values in plants is the type of soil-N used as a parameter (organic N, NH+ 4 , NO3), because ammonium tends to have higher d15N values than 15 + NO 3, and plants taking up NH4 will thus have higher d N [26] [27] values. Mariotti et al. reported that plants grown in N-decient soil have a higher capacity for uptake of NH+ 4 15 than NO N. 3, and are thus richer in Plant d15N values depend not only on the d15N values of the source nitrogen, but also on the balance between enzyme demand and external concentration.[28] In an open-pool system, a high external N concentration coupled with modest demand can inhibit 15N fractionation[29] if the rate of N uptake exceeds the assimilation differences between the d15N values of the plant and of the nitrogen in the soil.[43] Another factor that could directly affect stable nitrogen isotope ratios in cereals is manuring, since the application of animal dung raises d15N values in soil and plants.[3034] Analyzing the chemical composition of paleosoils could in turn provide direct evidence of manure input.[3537] Unfortunately, in arid and semi-arid regions the over-exploitation of the landscape makes it unlikely that such evidence is preserved.[3840] Observations made in the eld as well in the laboratory suggest that external nitrogen concentration is not the only factor that determines discrimination.[41,42] Indeed, the translocation of nitrogen to the site of assimilation can account for intra-plant patterns in d15N values. Signicant intra-plant variation can be observed when NO 3 is the primary nitrogen source, but little variation is observed when [43] NH+ 4 is the source. The use of the same plant components (e.g. kernels) can help to reduce the effects of different nitrogen uptake mechanisms and different assimilation and recycling pathways, but does not take account of the effects of genotype on 15N discrimination.[41,42,44] Despite the many studies dealing with the metabolic pattern in nitrogen in plants, it is still hard to determine the d15N values of external nitrogen sources and the physiological mechanisms within the plant which affect stable nitrogen isotope ratios.[43] Due to the large number of factors that may inuence stable nitrogen isotope ratios, d15N measurements are more useful when combined with d13C values, since the latter retain a clearer source signature.[44] Indeed, the use of such isotopes has been shown to be of some importance in ecological studies.[45] The long tradition of studies based on the use of stable carbon isotopes has demonstrated the relationship between d13C values in plants and local climate variables. Numerous investigations have shown that the d13C values of plants are negatively correlated with water input. This negative correlation is more pronounced in arid environments than in humid areas.[4649] A possible explanation lies in the assumption that water availability rather than other factors is the key climatic limiting factor for growth and survival in semiarid areas.[50] As a result, variations in precipitation, atmospheric humidity and soil water availability are assumed to account for changes in plant 13C concentration in sub-arid areas.[51] Thus, by combining stable carbon and nitrogen isotope analyses of charred grains of measured volume, we can identify differences between stored grain samples and determine whether they were harvested in different locations under Eblas control and whether irrigation or manuring was applied to make up for local resource decits.

THE CONTEXT OF THE STUDY


The early city-state of Ebla The Ebla archaeological site, with its long human history from the 3rd to the 2nd millennium BC, is one of the most important early historical settlements in north-western Syria.[52] Situated in a sub-arid region, but bordering a more humid and fertile ecosystem on which it relied, the settlements survival strictly depended on annual rainfall, which is nowadays around 450/500 mm.[53] The hydrography of the area is not suitable for ancient irrigation systems. In the surroundings of the site there are no big rivers which could provide additional water by means of channels, and no traces of subterranean tunnels for rainfall collection (qanat) have yet been found in the area.[54] The absence of any reference in the Ebla administrative texts to water management or controversies over this issue suggests that agriculture in the area was based on a rain-fed system. The overall pattern that emerges from the archaeological evidence and epigraphic sources points to an economy based on the farming of state-owned elds and a taxation system which provided enough foodstuffs to support the bureaucracy and the army. The storage system Eblas inuence extended all across north-western Syria. Raw materials were imported from distant locations while foodstuffs came from productive areas nearby. Cereals were harvested by local peasant communities based in numerous satellite settlements around Ebla, which paid a tribute to the palace in the form of produce. The palace ground the cereals and re-distributed them as rations of our or bread to the palace workers.[55] The storeroom (L.9512), sited in the B-east quarter, can be considered a place where foodstuffs were stored before being processed. The architectural characteristics of L.9512 show that it was not sealed,[56] so grain could not be stored for long periods.[57]

328

wileyonlinelibrary.com/journal/rcm

Copyright 2012 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2012, 26, 327335

Studying ancient crop provenance

EXPERIMENTAL
Grain analysis Fresh material Modern samples of barley (Hordeum vulgare ssp. vulgare and H. vulgare ssp. distichon) from 21 sites in Syria were selected from the ICARDA (International Centre for Agricultural Research in Dry Areas, Aleppo, Syria) landrace seed bank.a The selection of sites was designed to maximise environmental variability[58] while maintaining similar land management to that of ancient crops. Irrigated and fertilized elds thus were avoided (Fig. 1). Each collection site was characterized by caryopses with homogeneous dimensional traits; in order to determine the average volume of each population, the thickness, length and width of ve fresh grains were measured. They were then reduced to a ne powder using an agate pestle and mortar and analyzed by isotope ratio mass spectrometry (IRMS). The effect of charring was also tested on modern samples in order to verify their comparability with the ancient charred remains. The charring experiments involved placing a crucible full of sand with 5 grains of modern, air-dried barley wrapped in foil in a hot oven at 250 C for 12 h in an anaerobic atmosphere. These conditions were chosen on the basis of previous studies, which have shown that they provide the closest replication of ancient charring conditions.[5] The samples were then re-measured before being reduced to a ne powder using an agate pestle and mortar and subjected to IRMS analysis. Archaeological material Charred grains of ancient barley were recovered from the oor of storeroom L.9512 in the Ebla site during the 2005 2006 campaigns. The storeroom (L.9512) was a rectangular room of around 5 m2. Stratigraphical observations identied three layers (0-1-2) of charred caryopses separated from each other by debris. Since it was hypothesised that tributes of grain from different elds were stored together, we intended to identify correspondences between their spatial and vertical distribution and isotopic-volume signature. Specimens (5 L) were thus taken every 50 cm2; six samples were taken from each layer, making 18 in total. Once the samples had been obtained, a standard otation tank with a 0.5 mm sieve was used in the eld to separate plant remains from other remains. Afterwards, wet samples were dried slowly before being stored. In the upper layer (Level 0), as well as in the bottom layer (Level 2), caryopses were recovered in ve out of six samples, while in the intermediate layer (Level 1) only three of the six yielded caryopses. Thus, only thirteen specimens contained cereals. Seven caryopses of barley were randomly selected from each sample and their morphological features analyzed in detail by microscopy (Nikon SMZ645; Nikon Instruments, Tokyo, Japan). The thickness, length and width of the

grains were also measured and the average volume estimated assimilating the barley caryopses to an ellipsoid, as follows: L W T=2 4=3p (1)

The grain weight was estimated by using equations derived from the regression models of Ferrio et al.:[59] 0:21 L W 1:6 0:78 L T 1:28 (2) (3)

Carbonate was removed from the fossil grains by soaking each grain separately in 6 M HCl for 24 h at room temperature and then rinsing the grain repeatedly with distilled water.[60] Finally, the 91 caryopses were divided into 13 groups on the basis of their position in the storeroom and ground to a ne powder. Principal component analysis (PCA) was used to identify the presence of clusters among the variables (volume, and d13C and d15N values) and to obtain information about the importance of the variables. Stable isotope analysis Sample preparation and isotope measurements The isotope signature of fresh material was determined for 21 samples selected from ICARDA varieties, before and after charring treatment, making a total of 42 samples. For archaeological measurements, seven charred caryopses from each of the thirteen samples were analyzed. Approximately 0.5 mg of each sample was placed in tin capsules for measurement. The d13C and d15N values were measured by continuous ow elemental analysis isotope ratio mass spectrometry (CF-EA-IRMS) using a Delta V Advantage mass spectrometer (Thermo Fisher Scientic, Bremen, Germany) interfaced with a FlashEA 1112 HT elemental analyzer (Thermo Fisher Scientic) (oxidation/reduction reactor temperature: 1050 C and gas chromatography column temperature: 50 C). The CF-EA-IRMS system was operated in dual isotope mode, allowing d13C and d15N values to be measured in the same sample. Standards and equations The results of the carbon (d13C) and nitrogen (d15N) isotope ratio analyses were reported as per mil (%) relative to V-PDB (Vienna-Pee Dee Belemnite) and Atmospheric Air, respectively, after calculation in accordance with the following equation: d% RSample RReference =RReference 1000 where R is the ratio of the heavy to light stable isotopes in the sample (RSample) and the standard (RReference). Each sample was analyzed twice and the values were compared with those of reference gases (N2 and CO2) previously calibrated against International Standards specied by the International Atomic Energy Agency (IAEA, Vienna, Austria) (IAEA-NO3 for d15N values and IAEA-CH6 NBS 22 for d13C values). For each run at least one in-house casein standard was analyzed at regular intervals to check the accuracy of measurements in case of drift deviations.

Samples were not multiplied in ICARDA, but harvested directly from elds spreading across the Syrian territories.

329

Rapid Commun. Mass Spectrom. 2012, 26, 327335

Copyright 2012 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/rcm

G. Fiorentino et al.

Figure 1. Geological map of Syria and the site-sampling distribution. Numbers refer to sampling sites listed in Table 1. Statistical analysis Different statistical analyses were performed in order to: (i) ascertain the effect of charring on grain size, and the d13C and d15N values of modern plant material; (ii) establish the relationship between d13C values and grain size, and the carbon and nitrogen stable isotope values in the ancient material; and (iii) discriminate groups among the ancient stored cereals. The rst task was achieved by applying linear regression to estimate the correlation between fresh and charred volume, and fresh and charred d13C values, and to predict change in the d15N values after charring. Then, the degree of correlation between d13C values and size, and the nitrogen and carbon stable isotope ratios in the ancient material was determined by logarithm regressions. The last task (iii) was achieved by applying statistical analysis on the archaeobotanical data, and was performed using Statistica 8.0 (StatSoft Italia srl, Padova, Italy). PCA enabled us to visualize the presence of clusters of samples and to obtain information about the importance of variables (volume, and d13C and d15N values) in these groupings. The variables were scaled by using auto-scaling method, i.e. the variable mean was subtracted from each variable and then each variable was divided by its standard deviation. Analysis of variance (ANOVA) provided the statistical significance of each group. Unless otherwise stated, differences were considered statistically signicant when P <0.05. Previous studies have shown that the carbon isotope signature, which is strongly dependent on local climate conditions, varies from site to site but is fairly consistent across sites with similar ecological conditions. The average d13C value is about 24.5%, with a variation of ~1.52.0% between the samples from Naher Markia-Tartous and those collected at Kawkab-Hama. The variation in d15N values, of the order of ~5%, was greater because the d15N value primarily reects the isotope signature of the N source, which is not easy to determine, and discrimination processes which occur during N uptake, assimilation and redistribution within the plant. The effect of charring The charring experiments revealed that controlled carbonization determines an increase in caryopsis volume. Since this is fairly constant for all specimens, the volume of the charred barley was closely related to that of the fresh material (r2 = 0.59 with linear correlation). The estimated grain weights showed a lower degree of correlation than the volume: r2 = 0.53, Eqn. (2), r2 = 0.43, Eqn. (3) (results not shown). Our results are in agreement with previous studies carried out on barley, such as those quoted by Renfrew,[12] who reported that slight increases in the length and breadth of barley grains have a limited effect on the shape of each single grain. The effect of charring on the d13C and d15N values of grains has been widely explored (Fig. 2). Aguilera et al.[4] reported that d13C values in cereal grains are not signicantly affected by charring within the temperature range normally expected to induce carbonization and sample preservation (ca. 200400 C), and that the original d13C environmental signal is retained in charred grains. Heaton et al.[5] also found that differences in charring conditions and duration had little inuence on d13C values. In this study we found that no changes occurred in the d13C values as a result of carbonization. Indeed, the correlation between fresh and charred values was r2 = 0.87, with linear correlation (results not shown).

RESULTS
Estimation of volume and isotope signature in fresh material The results of the analyses carried out on the 21 ICARDA varieties are shown in Table 1. The volume measurements reveal signicant differences between different areas of Syria. With an average value of 37 mm3, the volume ranged from 25 mm3 for the samples collected at Baovar-Dar a to 55 mm3 for those gathered at Quadir-Homs.

330

wileyonlinelibrary.com/journal/rcm

Copyright 2012 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2012, 26, 327335

Table 1. Summary of volume, and d13C and d15N values before and after the combustion of the 21 modern barley land-races Volume (mm2) Lat N35 05 N35 16 20 N32 37 30 N33 13 00 N34 36 05 N 35 56 51 N36 24 02 N36 41 00 N36 50 30 N36 22 20 N34 44 09 N35 41 N35 02 00 N35 01 55 N35 34 50 N33 39 15 N34 34 10 N35 06 N35 17 40 N35 37 30 N32 49 20 E36 41 14 E38 45 20 E36 47 30 420 270 800 E38 54 470 E38 21 22 410 E36 50 10 655 E35 49 00 20 E35 54 50 25 E35 54 00 100 ssp. vulgare covar. ssp. vulgare covar. ssp. vulgare covar. ssp. vulgare covar. ssp. vulgare covar. ssp. vulgare covar. ssp. vulgare covar. ssp. vulgare covar. ssp. vulgare covar. E35 50 10 ssp. vulgare covar. E36 43 00 480 ssp. vulgare covar. 44.82 40.98 32.38 29.32 30.62 35.19 42.67 42.92 49.13 31.84 32.59 E37 31 49 470 ssp. vulgare covar. 44.18 E40 04 20 350 ssp. vulgare covar. 33.84 64.71 69.58 82.63 82.84 56.62 61.28 77.80 69.76 67.34 82.53 88.05 59.37 64.46 E39 56 00 410 ssp. vulgare covar. 33.54 55.06 E39 16 45 350 ssp. vulgare covar. 26.12 49.16 E39 02 10 250 35.66 52.25 E37 06 07 680 42.45 64.94 E36 25 00 650 H. vulgare ssp. vulgare 25.41 39.93 E36 13 30 600 H. vulgare ssp. vulgare 36.60 68.55 E38 51 00 460 H. vulgare ssp. vulgare 55.69 83.39 E38 55 470 H. vulgare ssp. vulgare 34.03 62.44 Long Alt Sample fresh charred fresh charred fresh d13C (%) (SD) d15N (%) (SD) charred

IG

Site

31395

Altaibe (Homs)

31397

Quadir (Homs)

Studying ancient crop provenance

31405

Sayda (Dar a)

31407

Baovar (Dar a)

331

Rapid Commun. Mass Spectrom. 2012, 26, 327335

31411

Furqlos (Homs)

31413

Raqqa

H. vulgare ssp. vulgare covar. distichon H. vulgare ssp. vulgare

31417

Raqqa (Kantari)

31421

31423

Al Hassake (Tell Hanish) Rasalain

10

31435

Al Bab (Aleppo)

11

31448

Homs

12

32709

13

32723

Jeb Al Kabier (Lattakia) Tartous

Copyright 2012 John Wiley & Sons, Ltd.

14

32735

15

32740

Naher Markia (Tartous) Kanjara (Lattakia)

16

128172

Awajan (Dams)

17

128173

Tadmor (Homs)

18

138175

Altaibe (Homs)

19

138176

Kawkab (Hama)

20

138220

Raqqa (Resafa)

21

139368

Hayran (Sweida)

H. vulgare vulgare H. vulgare distichon H. vulgare distichon H. vulgare distichon H. vulgare distichon H. vulgare Vulgare H. vulgare vulgare H. vulgare vulgare H. vulgare vulgare H. vulgare distichon H. vulgare distichon H. vulgare vulgare H. vulgare distichon H. vulgare distichon H. vulgare distichon

23.45 (0.03) 23.70 (0.03) 23.96 (0.03) 23.54 (0.05) 23.78 (0.01) 24.25 (0.02) 22.39 (0.04) 24.59 (0.02) 25.37 (0.01) 24.03 (0.04) 24.63 (0.02) 26.66 (0.01) 23.55 (0.02) 23.41 (0.02) 23.52 (0.01) 26.5 (0.02) 26.28 (0.03) 25.53 (0.03) 26.69 (0.03) 25.19 (0.01) 24.55 (0.02)

23.2 (0.03) 23.3 (0.02) 23.4 (0.05) 23.7 (0.01) 24.1 (0.03) 24.1 (0.03) 22.7 (0.06) 24.5 (0.04) 24.7 (0.08) 24.6 (0.05) 24.8 (0.02) 26.2 (0.01) 23.5 (0.03) 23.2 (0.02) 23.5 (0.02) 26.6 (0.03) 25.7 (0.01) 26 (0.04) 26.7 (0.07) 26.4 (0.03) 24.3 (0.05)

0.06 (0.02) 0.50 (0.01) 0.81 (0.02) 0.14 (0.02) 0.09 (0.01) 0.21 (0.01) 0.84 (0.02) 0.13 (0.03) 2.17 (0.01) 2.04 (0.02) 1.10 (0.02) 0.06 (0.01) 0.29 (0.02) 0.03 (0.03) 0.37 (0.03) 1.23 (0.03) 4.48 (0.03) 1.30 (0.02) 1.79 (0.01) 0.13 (0.01) 0.48 (0.02)

0.30 (0.02) 0.20 (0.02) 0.60 (0.01) 0.10 (0.03) 2.03 (0.01) 0.03 (0.01) 0.10 (0.04) 2.01 (0.02) 1.02 (0.03) 2.05 (0.01) 2.01 (0.02) 1.02 (0.02) 0.10 (0.06) 0.09 (0.07) 0.07 (0.01) 1.07 (0.02) 3.06 (0.03) 1.40 (0.05) 1.70 (0.09) 0.08 (0.03) 1.80 (0.02)

wileyonlinelibrary.com/journal/rcm

SD: standard deviation

G. Fiorentino et al. The average volumes ranged from 6.06 mm3 for sample 1 to 106 mm3 for sample 6. Intra-sample variability (expressed by the standard deviation) was quite low compared with the variability between samples. This is consistent with the hypothesis that the storeroom contained grain harvested from different sites. It is noteworthy that the volume of each sample seems to be directly related to the d13C values (Fig. 3) (r2 = 0.69). The stable carbon isotope d13C values ranged from 21.7% to 23.3%, and the stable nitrogen isotope d15N values from 3.1% to 8.0%. Notably, the d15N values are all positive, which means the nitrogen sources themselves had positive values.[61] The d13C and d15N values are inversely correlated, with low 13 d C values corresponding to high d15N values, and vice versa (Fig. 4) (r2 = 0.51). After applying PCA to the data, two PCs were extracted. The variance explained by PC1 and PC2 was 74.9% and 20.4%, respectively. The analysis identied three major groupings a, b and c (Fig. 5); the statistical signicances of volume, d13C value and d15N value are, respectively, P = 0.007, P = 0.02 and P = 0.02.

Figure 2. Correlation between the fresh and charred caryopses volume of modern specimens. Regarding the effects of charring on d15N values, Bogaard et al.[7] reported no signicant changes in d15N values for grain carbonized at 230 C for up to 24 h. The work of DeNiro and Hastorf[60] on the d15N values of charred seeds and tubers from Peru showed that charring did not affect the nitrogen isotope signature of ancient plant remains. The experiments carried out on Spanish wheat and barley by Aguilera et al.[4] revealed that although signicant changes in element concentration (%N) occur during carbonization, the d15N value is not altered by carbonization and environmental signals are well preserved in charred grains. In the present study we found slight increases in d15N values, which were fairly constant across all samples (r2 = 0.63, with linear correlation) (results not shown). The Ebla silo Size and isotope analyses of archaeobotanical remains The volume of the caryopses collected in the storeroom was determined and differences were found among the samples from the thirteen squares. For each of these thirteen samples considered, the average value of seven grains was calculated (Table 2).

DISCUSSION
The experiment performed on modern samples shows that charring does not alter the general trend of either isotope signatures or volumes. It can thus be assumed that the d13C and d15N values, and the size of the carbonized caryopses, provide information about plant growth environments and distinguish between crops grown under different climate conditions or by different farming methods. With the ancient material, the correlation between the volume and the d13C value of the Ebla silo caryopses is consistent with the hypothesis that grain size depends mainly on water input during grain lling. In contrast, the volume and stable nitrogen isotope ratios are negatively correlated. Moreover, previous experiments carried out on barley grown in controlled environments show that d15N usually increases in response to drought, while it decreases under N-starvation.[21]

Table 2. Summary of volume, and d13C and d15N values of the ancient barley found in the three layers of the silo L.9512 Average volume (mm3) (SD) 6.06 (1.3) 14.26 (2.1) 66.88 (5.4) 83.60 (7.6) 65.83 (5.0) 106.59 (9.6) 14.81 (1.9) 62.70 (4.8) 32.91 (3.2) 37.62 (2.8) 61.86 (4.1) 54.13 (3.8) 30.17 (2.5) d13C (%) (SD) 21.71 (0.08) 22.80 (0.10) 22.97 (0.14) 22.66 (0.12) 23.12 (0.09) 23.33 (0.05) 21.93 (0.05) 22.99 (0.04) 22.70 (0.03) 22.64 (0.11) 23.04 (0.02) 22.93 (0.02) 22.95 (0.03) d15N (%) (SD) 8.00 (0.01) 5.17 (0.21) 6.66 (0.20) 6.37 (0.01) 4.57 (0.09) 3.15 (0.40) 5.95 (0.13) 3.34 (0.31) 4.41 (0.07) 4.89 (0.08) 4.58 (0.04) 4.5 (0.04) 3.44 (0.31)

Layer 0

Group 1 2 3 4 5 6 7 8 9 10 11 12 13

Sample H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare H. vulgare

1 2

SD: standard deviation

332

wileyonlinelibrary.com/journal/rcm

Copyright 2012 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2012, 26, 327335

Studying ancient crop provenance higher d15N values. These may have less capacity to restrict organic-N loss from their roots when under stress than more stress-tolerant genotypes.[62] These experiments provide an interpretative basis for our analyses; indeed, the period in which the grain was stored (around 1800 BC) corresponds, on the basis of various proxies, to a drought crisis.[6466] Therefore, the inverse correlation between the d13C and d15N values in our ancient samples can be interpreted as the response of barley to environmental stress brought about by water scarcity.[67] On the other hand, we cannot exclude the possibility that manure was added by farmers in response to water scarcity, causing the d15N values to increase despite the low d13C values. However, we cannot exclude the possibility that differences in d15N values can depend on the genotype. Indeed, previous experiments carried out on 28 genotypes of Hordeum spontaneum showed that variation in whole-plant d15N values can reach 1.5%.[21] Variability in terms of volume (P = 0.007), d13C value (P = 0.02) and d15N value (P = 0.02) of barley is consistent with the hypothesis that the cereals were harvested from different elds. The possibility that the variability was due to temporal differences in harvesting can be excluded the since the administrative texts record the annual storage practices.[68]

Figure 3. Correlation between grain size and d13C value of archaeobotanical remains.

CONCLUSIONS
Figure 4. Correlation between d C and d N values of caryopses found in the silo.
13 15

The stable isotope measurements and size analyses of grain found in a storeroom open up new perspectives for research into the provenance and/or storage manners of ancient foodstuffs. This new archaeobotanical approach supports the hypothesis that isotope values and grain size mostly reect environmental conditions and/or agricultural practices. Further studies are required to disentangle the effects of agricultural practices (additional water input, manuring) from those of local climate conditions. Nonetheless, the study of silo L.9512 sheds light on the role of Ebla as the centre of a complex rural system, where raw materials harvested from different sites were stored together before being ground and redistributed as rations of our. Thanks to the innovative approach adopted in this study, the contents of the silo, apparently homogenous, were shown to be made up of grain from sites characterized by different growing conditions.

Figure 5. PCA analysis with the three groups highlighted.

Acknowledgements
The authors are grateful to Ian Walkoun and ICARDA G_Bank for providing the modern crop material from Syria. The study was partially funded by the IAM-CIHEAM International Cooperation Project called Rationalization of irrigation systems in Ras al Ain Syria. Special thanks to Paolo Matthiae and Luca Peyronel for allowing the study of the archaeobotanical material from Ebla and Fabio Mavelli for statistical discussions. The useful comments and suggestions of two anonymous referees helped us to improve the original text.

According to Robinson et al.[62] the genotypes which are most productive and stress tolerant are those which, when under stress, preferably restrict the loss of N from their roots 15 to ammonium (NH+ N, while retaining 4 ), which is richer in [63] organic nitrogen. Van der Leij et al. also found that NO3 exudation from the roots of N-starved plants is negligible. Conversely, genotypes characterized by more limited discrimination against 15N were smaller and had

333

Rapid Commun. Mass Spectrom. 2012, 26, 327335

Copyright 2012 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/rcm

G. Fiorentino et al.

REFERENCES
[1] T. H. E. Heaton. Spatial, species, and temporal variations in the 13C/12C ratios of 13C C3 plants: implications for palaeodiet studies. J. Arch. Sci. 1999, 26, 637. [2] G. Eriksson, A. Linderholm, E. Fornander, M. Kanstrup, P. Schoultz, H. Olofsson, K. Lidn. Same island, different diet: cultural evolution of food practice on land, Sweden, from the Mesolithic to the Roman period. J. Antr. Arch. 2008, 27, 520. [3] G. Fiorentino, V. Caracuta, L. Calcagnile, M. DElia, P. Matthiae, F. Mavelli, G. Quarta. Third millennium B.C. climate change in Syria highlighted by carbon stable isotope analysis of 14 C-AMS dated plant remains from Ebla. Paleogog., Palaeoclim., Palaeoecol. 2008, 266, 51. [4] M. Aguilera, J. L. Araus, J. Voltas, M. O. Rodrguez-Ariza, F. Molina, N. Rovira, R. Bux, J. P. Ferrio. Stable carbon and nitrogen isotopes and quality traits of fossil cereal grains provide clues on sustainability at the beginnings of Mediterranean agriculture. Rapid Commun. Mass Spectrom. 2008, 2, 1653. [5] T. H. E. Heaton, G. Jones, P. Halstead, T. Tsipropoulos. Variations in the 13C/12C ratios of modern wheat grain, and implications for interpreting data from Bronze Age Assiros Toumba, Greece. J. Arch. Sci. 2009, 36, 224. [6] J. L. Araus, A. Febrero, R. Bux, M. O. Rodrguez-Ariza, F. Molina, M. D. Camalich, D. Martn, J. Voltas. Identication of ancient irrigation practises based on the carbon isotope discrimination of plant seeds: a case study from the south east Iberian Peninsula. J. Arch. Sci. 1997, 24, 729. [7] A. Bogaard, T. H. E. Heaton, P. Poulton, I. Merbach. The impact of manuring on nitrogen isotope ratios in cereals: archaeological implications for reconstruction of diet and crop management practices. J. Arch. Sci. 2007, 34, 335. [8] H. V. Harlan. Barley: Culture, Uses, and Varieties, U.S. Department of Agriculture, Farmers Bulletin n 1464, 1932. [9] R. von Bothmer, N. Jacobsen, C. Baden, R. B. Jrgensen, I. Linde-Laursen. An Ecogeographical Study of the Genus Hordeum, (2nd edn.), Systematic and Ecogeographical Studies on Crop Genepools 7. International Plant Genetic Resources Institute, Rome, 1995. [10] G. Azzi. Le climat du bl dans le monde. Institut International dAgriculture, Rome, 1930. [11] J. Percival. Wheat in Great Britain. Duckworth, London, 1943. [12] J. M. Renfrew. Palaeoethnobotany. Methuen & Co., London, 1973. [13] S. Kelly, K. Heaton, J. Hoogewerff. Tracing the geographical origin of food: the application of multi-element and multi-isotope analysis. Trends Food Sci. Technol. 2005, 16, 555. [14] A. Sacco, M. A. Brescia, A. Sgaramella, D. Sacco. Characterization of the composition and the geographical origin of food products by means of nuclear magnetic resonance and isotope ratio mass spectrometry. Recent Res. Develop. Agric. Food Chem. 2005, 6, 119. [15] J. Comadran, J. R. Russell, et al. Mapping adaptation of barley to droughted environments. Euphytica 2008, 161, 35. [16] P. H. Beatty, Y. Anbessa, P. Juskiw, R. T. Carroll, J. Wang, A. G. Good. Nitrogen use efciencies of spring barley grown under varying nitrogen conditions in the eld and growth chamber. Ann. Bot. 2010, 105, 1171. [17] S. Ceccarelli, S. Grando. Selection environment and environmental sensitivity in barley. Euphytica 1991, 57, 157. [18] A. Jilal, S. Grando, R. J. Henry, L. S. Lee, N. Rice, H. Hill, M. Baum, S. Ceccarelli. Genetic diversity of ICARDAs worldwide barley landrace collection. Gen. Resour. Crop Evol. 2008, 55, 1221. [19] S. Branch, S. Burke, P. Evans, B. Fairman, C. S. L. Wolff Briche. A preliminary study in determining the geographical origin of wheat using isotope ratio inductively coupled plasma mass spectrometry with 13C, 15N mass spectrometry. J. Anal. At. Spectrom. 2003, 18, 17.

[20] J. Voltas, I. Romagosa, J. L. Araus. Grain size and nitrogen accumulation in sink-reduced barley under Mediterranean conditions. Field Crops Res. 1997, 52 , 117. [21] D. Robinson, L. L. Handley, C. M. Scrimgeour, D. C. Gordon, B. P. Forster, R. P. Ellis. Using stable isotope natural abundances (d15N and d13C) to integrate the stress responses of wild barley (Hordeum spontaneum C. Koch.) genotypes. J. Exp. Bot. 2000, 51, 41. [22] N. E. West, J. Skujins. Nitrogen cycle in North American cold-winter semi-desert ecosystems. Oecol. Plant. 1977, 12, 45. [23] W. T. Peterjohn, W. H. Schlesinger. Factors controlling denitrication in a Chihuahuan desert ecosystem. Soil Sci. Soc. Am. J. 1991, 55, 1694. [24] K. J. Nadelhoffer, B. Fry. Nitrogen isotope studies in forest ecosystems, in Stable Isotopes in Ecology and Environmental Science, (Eds: K. Lajtha, R. Michener). Blackwell Scientic Publications, Boston, USA, 1994, p. 23. [25] R. D. Evans, A. J. Bloom, S. S. Sukrapanna, J. R. Ehleringer. Nitrogen isotope composition of tomato (Lycopersicon esculentum Mill. cv. T-5) grown under ammonium or nitrate nutrition. Plant Cell Environ. 1996, 19, 1317. [26] P. Hgberg. Tansley d15N natural abundance in soil-plant systems, Review No. 95. New Phytol. 1997, 137, 179. [27] A. Mariotti, F. Mariotti, N. Amarger, G. Pizelle, J. M. Ngambi, M. L. Champigny, A. Moyse. Fractionnement isotopiques de lazote lors des processus dabsorption des nitrates et de xation de lazote atmosphrique par les plants. Phys. Veg. 1980, 18, 163. [28] A. D. M. Glass, J. E. Shaff, L. V. Kochian. Studies of the uptake of nitrate in barley. Plant Physiol. 1992, 99, 456. [29] R. Bol, J. Eriksen, P. Smith, M. H. Garnett, K. Coleman, B. T. Christensen.The natural abundance of 13C, 15N, 34S and 14C in archived (19231996) plant and soil samples from the Askov long-term experiments on animal manure and mineral fertiliser. Rapid Commun. Mass Spectrom. 2005, 19, 3216. [30] C. M. Cossani, G. A. Slafer, R. Savin. Co-limitation of nitrogen and water, and yield and resource-use efciencies of wheat and barley. Crop Pasture Sci. 2010, 61, 844. [31] T. Yoneyama, K. Kouno, J. Yazaki. Variation of natural 15N abundance of crops and soils in Japan with special reference to the effect of soil conditions and fertilizer application. Soil Sci. Plant Nutr. 1990, 36, 667. [32] G. H. Wagner. Stable Isotopes in Plant Nutrition, Soil Fertility and Environmental Studies. International Atomic Energy Agency, Vienna, 1991. [33] I. A. Simpson, R. Bol, I. D. Bull, R. P. Evershed, K. J. Petze. Compound-specic stable isotope signals in anthropogenic soils as indicators of early land management. Rapid Commun. Mass Spectrom. 1999, 13, 1315. [34] W. J. Choi, S. M. Lee, H. M. Ro, K. C. Kinm, S. H. Yoo. Natural 15N abundances of maize and soil amended with urea and composted pig manure. Plant Soil 2002, 245, 223. [35] M. G. Canti. An investigation of microscopic calcareous spherulites from herbivore dungs. J. Arch. Sci. 1997, 24, 219. [36] R. P. Evershed, P. H. Bethell, P. J. Reynolds, N. J. Walsh. 5b-Stigmastanol and related 5b-stanols as biomarkers of manuring: analysis of modern experimental material and assessment of the archaeological potential. J. Arch. Sci. 1997, 24, 485. [37] E. B. A. Guttman. Midden cultivation in prehistoric Britain: arable crops in gardens. World Arch. 2005, 37, 224. [38] R. G. Amundson. The use of stable isotopes in assessing the effect of agriculture on arid and semi-arid soils, in Stable Isotopes in Ecological Research, (Eds: P. W. Rundel, J. R. Ehleringer, K. A. Nagy). Ecological Studies 68. SpringerVerlag, New York, 1988, p. 318.

334

wileyonlinelibrary.com/journal/rcm

Copyright 2012 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2012, 26, 327335

Studying ancient crop provenance


[39] R. D. Evans, J. R. Ehleringer. A break in the nitrogen cycle in arid lands? Evidence from 15N of soils. Oecologia 1993, 94, 314. [40] R. Shahack-Gross, A. Simons, S. H. Ambrose. Identication of pastoral sites using stable nitrogen and carbon isotopes from bulk sediment samples: a case study in modern and archaeological pastoral settlements in Kenya. J. Arch. Sci. 2008, 35, 983. [41] A. Mariotti. Mmoire des Science de la Terre, Universit P. et M. Curie. Paris, 1982. [42] A. Mariotti, F. Mariotti, M. L. Champigny, N. Amarger, A. Moyse. Reductase activity and uptake of NO 3 by pearl millet. Plant Phys. 1982, 69, 880. [43] R. D. Evans. Physiological mechanisms inuencing plant nitrogen isotope composition. Trends Plant Sci. 2001, 6, 121. [44] L. L. Handley, J. A. Raven. The use of natural abundance of nitrogen isotopes in plant physiology and ecology. Plant Cell Environ. 1992, 15, 965. [45] H. Grifths. Applications of stable isotope technology in physiological ecology. Funct. Ecol. 1991, 5, 254. [46] R. L. Korol, M. U. F. Kirschbaum, G. D. Farquhar, M. Jeffreys. Effects of water status and soil fertility on the C-isotope signature in Pinus radiata. Tree Phys. 1999, 19, 551. [47] J. M. Miller, R. J. Williams, G. D. Farquhar. Carbon isotope discrimination by a sequence of Eucalyptus species along a sub-continental rainfall gradient in Australia. Funct. Ecol. 2001, 15, 222. [48] P. K. Van de Water, S. Leavitt, J. L. Betancourt. Leaf d13C variability with elevation, slope aspect and precipitation in southwest United States. Oecologia 2002, 132, 332. [49] S. Chen, Y. Bai, G. Lin, X. Han. Variations in life-form composition and foliar carbon isotope discrimination among eight plant communities under different soil moisture conditions in the Xilin River Basin, Inner Mongolia, China. Ecol. Res. 2005, 20, 167. [50] R. Nemani, C. D. Keeling, H. Hashimoto, W. M. Jolly, S. C. Piper, C. J. Tucker, R. B. Myneni, S. W. Running. Climatedriven increases in global terrestrial net primary production from 1982 to 1999. Science 2003, 300, 1560. [51] J. R. Ehleringer. Physiological processes in aridland plants, in Stable Isotopes in Ecological Research, (Eds: P.W. Rundel, J.R. Ehleringer, K. A. Nagy), Ecological Studies 68. SpringerVerlag, New York, 1988, p. 41. [52] P. Matthiae, Ebla. Un impero ritrovato. Dai primi scavi alle ultime scoperte. Einaudi, Torino, 1989. [53] M. Zohary. Geobotanical Foundations of the Middle East. Fisher Verlag, Stuttgart, 1973. [54] D. R. Lightfoot. Syrian Qanat Romani: History, Ecology, Abandonment. J. Arid. Environ. 1996, 33, 321. [55] L. Milano. Barley for ratio and barley for sowing (ARET II 51 and related matters). Acta Sum. 1987, 9, 177. [56] L. Peyronel. Domestic quarters, refuse pits and working areas. Reconstructing human landscape and environment at Tell Mardikh-Ebla during the old Syrian period (c. 2000 1600 BC), in Proc. 4th Int. Congress of the Archaeology of the Ancient Near East, vol. 1, (Eds: H. Khne, R.M. Czichon, F. J. Kreppner), 2008, p. 177. [57] Les Techniques de Conservation des Grains Long Terme, (Eds: M. Gast, F. Sigaut), Centre Nationale de la Recherche Scientique, Paris, 1979. [58] Diversity in Barley (Hordeum vulgare), (Eds: R. von Bothmer, T. Van Hintum, H. Knpffer, K. Sato), Elsevier Science B.V., Amsterdam, 2003. [59] J. P. Ferrio, N. Alonso, J. Voltas, J. L. Araus. Estimating grain weight in archaeological cereal crops: a quantitative approach for comparison with current conditions. J. Arch. Sci. 2004, 31, 1635. [60] M. J. DeNiro, C. A. Hastorf. Alteration of 15N/14N and 13 C/12C ratios of plant matter during the initial stages of diagenesis: Studies utilizing archaeological specimens from Per. Geochim. Cosmochim. Acta 1985, 49, 97. [61] J. Bort, J. L. Araus, H. Hazzam, S. Grando, S. Ceccarelli. Relationships between early vigor, grain yield, leaf structure and stable isotope composition in eld grown barley. Planta Phys. Biochem. 1998, 38, 889. [62] D. Robinson, L. L. Handley, C. M. Scrimgeour. A theory for 15 N/14N fractionation in nitrate-grown vascular plants. Planta 1998, 205, 397. [63] M. Van der Leij, S. J. Smith, A. J. Miller. Remobilization of vacuole stored nitrate in barley root cells. Planta 1998, 205, 64. [64] N. H. Dalfes, G. Kukla, H. Weiss. Third Millennium BC Climate Change and Old World Collapse, NATO ASI Series vol. 49, 1993. [65] S. Riehl, R. Bryson, K. Pustovoytov. Changing growing conditions for crops during the Near Eastern Bronze Age (30001200 BC): the stable carbon isotope evidence. J. Arch. Sci. 2007, 20, 1. [66] N. Roberts, W. J. Eastwood, C. Kuzucuolu, G. Fiorentino, V. Caracuta. Climatic, vegetation and cultural change in the eastern Mediterranean during the mid-Holocene environmental transition. Holocene 2011, 21, 147. [67] L. L. Handley, D. Robinson, B. P. Forster, R. P. Ellis, C. M. Scrimgeour, D. C. Gordon, E. Nevo, J. A. Raven. Shoot d15N correlates with genotype and salt stress in barley. Planta Rapid Commun. 1997, 201, 100. [68] R. Dolce. Some aspects of the primary economic structures of Ebla in the Third and Second Millenium BC: stores and working places, in Akten der Internationalen Tagung: Wirtschaft und Gesellschaft von Ebla (Eds: H. Waetzoldt, H. Hauptmann), Heidelberg. 1988, p. 35.

335

Rapid Commun. Mass Spectrom. 2012, 26, 327335

Copyright 2012 John Wiley & Sons, Ltd.

wileyonlinelibrary.com/journal/rcm

You might also like