Está en la página 1de 5

RAPID COMMUNICATIONS IN MASS SPECTROMETRY Rapid Commun. Mass Spectrom.

13, 13151319 (1999)

Interpreting Early Land Management Through Compound Specic Stable Isotope Analyses of Archaeological Soils
I. A. Simpson1*, R. Bol2, I. D. Bull3, R. P. Evershed3, K.-J. Petzke4 and S. J. Dockrill5
1 2

Department of Environmental Science, University of Stirling, Stirling FK9 4LA, Scotland, UK Institute of Grassland and Environmental Research, North Wyke, Okehampton EX20 2SB, UK 3 Organic Geochemistry Unit, School of Chemistry, University of Bristol, Bristol BS8 1TS, UK 4 German Institute for Human Nutrition, Arther-Scheunert-Allee 114116, D-14558 Bergholz-Rehbrucke, Germany 5 Department of Archaeological Sciences, University of Bradford, Bradford BD7 1DP, UK

Compound specic stable isotope analyses of managed soils using isotope ratio mass spectrometry have been undertaken as a means of determining early land use practices. d 15N amino acid signals demonstrate differences between manured grassland, unmanured grassland and continuous cereal cultivation under long-term experimental land use control conditions, with d 15N in hydrophobic amino acids providing the most distinctive signals. Analysis of early modern/medieval and of Bronze age anthropogenic soils from Orkney demonstrates that such signals are retained in archaeological contexts. d 13C analyses of n- alkanoic acid components of the fossil, Bronze Age, anthropogenic soils suggest a major terrestrial input to these soils, with uniform composition of formation materials. Surcial soils demonstrate the assimilation of isotopically lighter carbon, providing a means of assessing the mobility of the n- alkanoic acids within soils and sediments. Copyright # 1999 John Wiley & Sons, Ltd.
Received 26 January 1999; Revised 14 April 1999; Accepted 23 April 1999

The identification of early land use and associated land management practices remains one of the key objectives of geo-archaeological investigation and is central to discussions of cultural landscape change and palaeo-economic interpretation. Organic geochemical studies of palaeosols around archaeological sites have made an increasing contribution to these objectives, with analyses of amino acid and lipid organic fractions holding particular promise. Quantitative distributions of soil amino acids with discriminant analysis classification procedures have been used to identify land use in experimental and archaeological contexts,14 while free soil lipids have been used to identify different sources of organic manure applied to early arable soils.5 However, while continuing to hold promise, the results from such analyses remain equivocal, lack validation in historical landscape contexts and require process based explanations. One way of taking organic geochemical analyses of archaeological soils further is through new sub-molecular analytical capabilities that identify compound specific stable isotope signals. If different land use activities give rise to variation in soil biochemical processes, then this will be reflected in distinctive compound specific stable isotope signals because of kinetic effects that discriminate between lighter and heavier isotopes.6 In this paper we empirically test this hypothesis by first identifying compound specific  15N amino acid signals in soils that have been subject to long-term experimental land use control. However, even if such distinctions exist under contemporary experimental conditions, it is necessary to test their preservation under the longer time periods involved in
*Correspondence to: I. A. Simpson, Department of Environmental Science, University of Stirling, Stirling FK9 4LA, Scotland, UK. CCC 09514198/99/13131505 $17.50

archaeological analysis and within the context of experimental control. To do this we examined compound specific  15N amino acid signals from relict (surface) and fossil (buried) anthropogenic soil profiles in Orkney. The relict soils date from the mediaeval period and the fossil soils from the Bronze age, giving two different time periods with which to identify compound specific  15N amino acid signals and compare them with the data derived from the experimental locations. In a further development, the paper goes on to introduce and assess the application of  13C analysis of n-alkanoic acids as a means of identifying organic manure sources in the Bronze age fossil soils. METHODS Soil samples Soil samples were collected from experimental farms where land use activities had remained constant for a number of decades. Surface soil samples (015 cm) from manured grassland and unmanured grassland came from the Palace Leas plots at Cockle Park experimental farm in Northumberland, England (55 13'N 1 41'W). Management of these plots has been unchanged since 1897 with manure originating from the experimental farm. Soil samples from a field maintained in cereal cultivation since 1972 through conventional tillage with no manure added came from the Boigneville experimental site in the Parisian basin France (48 20'N 2 20@E). Manure samples (1 to 3 days old) from the Institute of Grassland and Environmental Research (IGER) North Wyke experimental Rowden farm provided further contemporary data. Samples from a relict anthropogenic soil, formed between the late 12th/early 13th centuries and the late 19th century
Copyright # 1999 John Wiley & Sons, Ltd.

1316

STABLE ISOTOPE ANALYSIS OF ARCHAELOGICAL SOILS

AD, came from Netherskaill, Marwick, West Mainland Orkney (59 10'N 3 30'W) with samples collected from a freshly exposed profile at 20, 40 and 60 cm depths.7 Grass has covered this soil for at least the last 30 years. Samples from fossil anthropogenic soils, buried beneath calcareous wind blown sands and formed between ca. 1685 and 993 BC, came from Tofts Ness, Sanday, Orkney (59 13'N 2 70@W). For  15N analyses samples were collected from a freshly exposed profile beyond the early settlement site of Mound 8 (Royal Commission on the Ancient and Historic Monuments of Scotland, RCAHMS notation) at 6065, 70 75 and 7883 cm depths.5 For  13C analyses samples were collected at 05, 3336, 3942, 4548 and 5154 cm depths from a freshly exposed profile associated with settlement Mound 11. Compound specic d
15 15

a  15N value of 0%. Further details of the sample preparation and analysis can be found in Metges et al.8 Compound specic d acids
13

C determination of n-alkanoic

N amino acid determination

 N amino acid analyses were undertaken at the German Institute for Human Nutrition. Finely ground air-dried soil samples (0.250.5g) of the `2 mm fraction were subjected to conventional acid hydrolysis (6N HCl 24 h at 110 C) and neutralised with ammonia to pH 7.0. Each sample was then filtered under vacuum before passing it slowly through a chromatography column filled with cation exchange resin Dowex 50W (1640 mesh size) allowing extraction of acidic, basic and neutral amino acids. Once all the sample was loaded, the bound material was eluted with 2M NH4OH (pH 14). Samples were concentrated by rotary evaporation and stored at below 0 C prior to analysis. Amino acids were derivatised to n-pivaloyl-i-propyl amino acid esters through treatment with thionyl chloride solution in i-propanol and heating for 60 min at 100 C. The product was dried and dissolved in pyridine. After adding pivaloyl chloride the solution was acylated 30 min at 60 C and dichloromethane added after cooling. The mixture was then passed through a silica gel column and the eluate dried in a gentle nitrogen stream and redissolved in ethyl acetate for injection. Isotopic analysis was carried out on a Finnigan delta S isotope ratio mass spectrometer (Finnigan MAT, Bremen, Germany) coupled on line with a gas chromatograph (GC, Hewlett Packard 5890, Palo Alto, CA, USA) via a combustion interface generating and purifying N2 gas from GC separated compounds introduced into the isotope ratio mass spectrometer. The interface consisted of a combustion furnace reactor filled with copper oxide and platinum (980 C) and a reduction furnace filled with elemental copper (600 C). The introduction of a standard N2 gas of known isotopic composition at particular timepoints during the gas chromatographic run was used for calibration of the sample amino acid nitrogen. An Ultra 2 capillary column (50 m 0.32 mm i.d., 0.5 mm film thickness Hewlett Packard) with a carrier gas (helium) stream of 1 mLmin1 (head pressure 18 psi) was used for separation of the amino acids. A volume of 0.5 mL was injected splitless by autosampler. The injector temperature was 280 C and the following oven temperature gradient was used: 70 C, held for 1 min; 70220 C, ramp 3 C min1; 220300 C, ramp 10 C min1, held 8 min. In the range of natural abundances (0.3626 to 0.3736 at.% 15N the [15N]/[14N] ratio is usually expressed as  15N% = {Rsample/Rstandard 1} 103, where R is the [15N]/[14N] ratio. [15N]/[14N] ratios are derived from respective ratios of m/z 29 to 28 ion current signals of the mass spectrometer. The international standard for nitrogen is air (0.3663 at.% 15N) which has been assigned
Rapid Commun. Mass Spectrom. 13, 13151319 (1999)

 13C determinations of n-alkanoic acids were carried out at the School of Chemistry, University of Bristol. Air-dried samples were partially crushed with a pestle and mortar and sieved to 75 mm. All samples were Soxhlet extracted over 24 h using a dichloromethane/acetone (9:1 v/v) solvent system to obtain a total lipid extract (TLE). The TLEs were further fractionated into two fractions using an aminopropyl bond elute cartridge. The first fraction comprised neutral lipids whereas the second contained predominantly n-alkanoic acids which were converted to their corresponding methyl esters by reaction with a BF3methanol complex at 70 C for 10 min. Gas chromatography combustion/isotope ratio mass spectrometry (GCC/IRMS) was carried out on 1 mL sample aliquots using a Varian 3400 gas chromatograph (Varian associates, Walnut Creek, California, USA) fitted with a BP1 fused silica capillary column (50 m 0.32 mm i.d.) 0.2 mm film thickness, SGE, He carrier gas, fitted with a septum equipped temperature programmable injector (SPI) coupled to a Finnigan MAT Delta S stable isotope mass spectrometer, electron ionisation, 10 eV electron voltage, 1 mA electron current, 3 Faraday cup collectors masses 44, 45 and 46, CuO/Pt combustion reactor set to a temperature of 850 C. RESULTS AND DISCUSSION d
15

N values of amino acids

The land management control data demonstrate that distinctions in manuring and vegetation cover can be made by  15N values in the hydrophobic soil amino acids; hydrophilic amino acids provide no distinctions in land management practices (Fig. 1). Where grassland soils have not been manured valine (Val), and to a lesser extent alanine (Ala), leucine (Leu) and isoleucine (Ile) are isotopically lighter in comparison to manured areas, with differences of between 7.3 and 1.8%. Comparison of soil amino acid  15N values under unmanured grassland and unmanured cereals give distinctions in phenylalanine (Phe), and to a lesser extent threonine (Thr), that can be attributed to differences in land use independent of manuring practice.  15N values are lighter under cereals and heavier under grassland with differences of 7.8% in Thr and major differences of more than 31.2% in Phe. The rapidly decomposing nature of the amino acid fraction in soils9 has resulted in a partial loss of signal in the archaeological soil context. Nevertheless  15N signals have been retained in both the relict mediaeval/early modern soils and in the fossil Bronze age soils from Tofts Ness, allowing preliminary interpretations to be made (Fig. 1). Comparison of the control data set with the amino acid  15N values of Ala, Val, Leu and Ile from the Tofts Ness soils supports the view that these soils have been heavily manured; it is not however possible to determine the origin of the manure from the existing data set. Land use distinctions can also be inferred from the amino acid  15N signals; Phe and Thr both demonstrate distinctly lighter values corresponding to the lighter values of unmanured
Copyright # 1999 John Wiley & Sons, Ltd.

STABLE ISOTOPE ANALYSIS OF ARCHAELOGICAL SOILS

1317

Figure 1. Compound specific soil amino acid 

15

N values.

cereal production observed in the control data. Similar signals are also evident in the mediaeval/early modern soils suggesting an earlier cereal production activity followed by grassland cover. A process based explanation for these empirical results is required, particularly for Phe and Thr given their potential to distinguish land cover types, but is hampered by an absence of compound specific  15N analyses of amino acids and their pathways within soils. One explanation is that under cereals more Phe and Thr precursor is produced with consequent higher enzymatic fractionation, resulting in the lighter  15N values observed for Phe and Thr. This is supported by the  15N differences between Phe, Thr and their respective glutamic acid (Glu) and aspartic acid (Asp) sources, which are markedly wider under cereals in
Copyright # 1999 John Wiley & Sons, Ltd.

comparison to grassland. Within the control data the  15N difference between Glu and Phe is 37.5% for cereals and 4.7% for unmanured grassland;  15N differences between Aspx and Thr are 11.2% for cereals and 4.0% for unmanured grassland. This process could be driven by differences in nitrogen, phosphorus and potassium (NPK) inputs as indicated by Tanacs et al.,10 or by the amount of available light which is known to mediate in the production of Phe. Such fractionation effects could be further reinforced by isotopic fractionation relations between substrate and soil microorganisms. Experimental analyses have demonstrated that procaryotic organisms will fractionate isotopes uniquely when grown on different substrates;11,12 substrate distinctions under grass and cereal land cover are likely.
Rapid Commun. Mass Spectrom. 13, 13151319 (1999)

1318

STABLE ISOTOPE ANALYSIS OF ARCHAELOGICAL SOILS

Figure 2. 

13

C values of n- alkanoic acid homologues extracted from soils taken from around settlement Mound 11. The soil depth at which the samples were taken is indicated.

Observed fractionation effects may arise because Thr and Phe are more easily recycled within the soil than their respective sources. This explanation is unlikely however as Thr has extra functional groups and Phe is aromatic, suggesting that they are likely to be strongly bound to refractive organics in the soil. Alternatively, the observation that amino acids are heavier under unmanured pasture compared to cereals may reflect the greater microbial activity under perennial grass cover.13 This may result in larger losses of light N as gaseous products of denitrification under grass compared to cereal cover. The observations may also reflect differences in the pattern of organic deposition either by leakage from live roots or in the form of sudden root death with subsequent microbial induced decomposition that occurs following cereals being harvested13 d
13

C values of n- alkanoic acids

Initial inspection reveals all values lie within the range 27 to 37% (Fig. 2) with homologues becoming progressively depleted in 13C in direct relation to chain length (Fig. 2). The majority of  13C values obtained for any, single homologue all lie within a very narrow range. A notable
Rapid Commun. Mass Spectrom. 13, 13151319 (1999)

exception to this pattern is the 05 cm surficial soil sample taken from soils around Mound 11 at Tofts Ness. The majority of n-alkanoic acid homologues are depleted by ca. 1.5% relative to the isotopic composition of identical homologues obtained from samples in the buried anthropogenic soil. Interestingly, the C16 and C18 components are consistently 3 to 6% less depleted in 13C than the C20 homologue; a significantly larger difference than is observed between consecutive higher homologues. These observations suggest a major terrestrial input to the Tofts Ness fossil anthropogenic soils. The narrow range of  13C values between soils obtained for each homologue is consistent with the idea of a fairly uniform composition of formation materials. The depletion of 1.5%, exhibited by homologues from the surficial soil may be attributed to the fixing of isotopically lighter CO2 by contemporary components. This result holds particular importance regarding the question of mobility on n-alkanoic components in the soil environment, inferring little, or no, detectable vertical migration. Furthermore, this result also supports the notion that lipids in the deeper fossil soils are indeed ancient in origin and not derived from modern day sources.
Copyright # 1999 John Wiley & Sons, Ltd.

STABLE ISOTOPE ANALYSIS OF ARCHAELOGICAL SOILS

1319

CONCLUSIONS The results of this research are preliminary; replicate analyses and further control samples are required to assess the degree of variation within the palaeosol contexts discussed. Nevertheless, the results do indicate that compound specific  15N amino acid signals in palaeosols may be used to discriminate between the early land use activities superimposed upon these soils and that compond specific  13C analyses may contribute to the identification of manuring practice and pedogenic processes. Furthermore, independent historical and archaeological data support the observations made. The mediaeval/early modern anthropogenic soil from Marwick, Orkney is known to be associated with the area of land known as the tunmal which was permanently attached to the farmstead, received all the best manure and was heavily cultivated for cereal production.1417 Recent and current farming practice has seen a shift towards livestock production and this area of anthropogenic soil has been under grassland for ca. 30 years. At Tofts Ness independent analyses of the fossil soils around Mound 8 and around the nearby Mound 11 (RCAHMS notation) suggests a diverse source of amendments, including domestic midden and hearth residues derived from burning combinations of turf dung and seaweed.18 Cultivation is physically attested by evidence for the use of the ard in the Mound 11 sequences as well as numerous cultivation tools, many of which show signs of wear, identified within the coarse artefact assemblage. Furthermore, clear contrasts in molluscan assemblages between the fossil anthropogenic soils and soils from the wider fossil landscape have been used to suggest cereal cultivation.18,19 The results therefore are promising and further development of these techniques may provide an effective method of measuring the extent of manured cereal cultivation areas against pasture areas. Such information is particularly vital in modeling the economic potential of prehistoric settlement sites.

Acknowledgements
Thanks are due to Dr R. Shiel for the Palace Leas samples and to Dr. S Chenu for the Boigneville samples. The use of the NERC Mass Spectrometry facilities (Grants GR3/2951, GR3/3758, FG6136/01) is gratefully acknowledged.

REFERENCES
1. F. J. Stevenson, Humus Chemistry, Wiley, New York (1982). 2. J. Beavis, Amino acids in buried soils, in N. R. J. Fieller, D. D. Gilbertson and N. G. A. Ralph (Eds), Palaeoenvironmental Investigations, Oxford: BAR International Series 258, p. 113 124 (1985). 3. J. Beavis and G. MacLeod, (1996). Amino acid studies, in M. Bell, P. J. Fowler and S. W. Hillson (Eds), The Experimental Earthwork Project 19601992, CBA Research Report 100, Council for British Archaeology, York, pp. 121126 (1996). 4. J. Beavis and C. J. B. Mott, Geoderma 72, 259 (1996). 5. I. D. Bull, I. A. Simpson, P. F. van Bergen and R. P. Evershed, Antiquity 73, 86 (1999). 6. B. J. Peterson and B. Fry, Ann. Rev. Ecology and Systematics 18, 293 (1987). 7. I. A. Simpson, Journal of Archaeological Science 24, 365 (1997). 8. C. C. Metges, K.-J. Petzke and U. Hennig, J. Mass Spectrom. 31, 367 (1996). 9. H. W. Hunt, Ecology 58, 469 (1977). 10. L. Tanacs, T. Bartok, J. Matuz, K. Kovacs, L. Gero and I. Harmati, Cereal Research Communications 20, 257 (1992). 11. S. A. Macko and M. L. F. Estep, Org. Geochem. 6, 787 (1984). 12. S. J. Macko, M. L. Fogel (Estep), P. E. Hare and T. C. Hoering, Chemical Geology (Isotopic Geoscience Section) 65, 79 (1987). 13. D. W. Hopkins, Personal Communication. 14. I. A. Simpson, Scottish Geographical Magazine 109, 4 (1993). 15. I. A. Simpson, Scottish Geographical Magazine 110, 100 (1994). 16. W. P. L. Thomson, History of Orkney, Mercat Press, Edinburgh (1987). 17. A. Fenton, The Northern Isles: Orkney and Shetland. John Donald, Edinburgh (1978). 18. S. J. Dockrill, J. M. Bond, A. Milles, I. A. Simpson and J. Ambers, Tofts Ness, Sanday, Orkney: an integrated study of a buried Orcadian landscape, in R. Luff and P. Rowley-Conwy (Eds), Whither Environmental Archaeology? Oxbow Monograph 38, pp. 115132 (1994). 19. A. Milles, (1991). The Molluscan Biostratigraphy and Archaeology of Holocene Coastal Blown Sands in the British Isles, Unpublished PhD Thesis, University of Wales College at Cardiff.

Copyright # 1999 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 13, 13151319 (1999)

También podría gustarte