Documenta Praehistorica XXXIX (2012)
Pots and lipids: molecular and isotope evidence of food processing at Maharski prekop
Nives Ogrinc1, Marinka Gams Petriši;1, Dušan ?igon1, Andreja ?ibrat Gašpari;2 and Mihael Budja2
1 Department of Environmental Sciences, Jo/ef Stefan Institute, Ljubljana, SI nives.ogrinc@ijs.si< marinka.petrisic@ijs.si< dusan.zigon@ijs.si 2 Department of Archaeology, Faculty of Arts, University of Ljubljana, SI andreja.zibrat.gasparic@gmail.com< miha.budja@ff.uni-lj.si
ABSTRACT - The pottery assemblage from the Maharski prekop site was analysed to obtain insights into vessel use and husbandry practices. Total lipid extracts of pottery samples were subjected to gas chromatography (GC), gas chromatography-mass spectrometry (GC-MS), gas chromatography-com-bustion-isotope ratio mass spectrometry (GC-C-IRMS) and soft ionisation electrospray mass spectro-metric techniques ESI Q-TOF MS and ESI Q-TOF MS/MS. The charred organic deposits on vessels were AMS14C dated. The results show that some vessels were used for cooking ruminant meat, while in other traces of mixed non-ruminant and ruminant meat or plants and animal meat cooking were identified. Some vessels were used for milk processing.
IZVLEČEK - Analizirali smo ostanke keramičnih posod z Maharskega prekopa, da bi dobili dodatne informacije o uporabi keramičnih posod in zgodnji živinoreji. Lipide, ekstrahirane iz ostankov keramičnih posod, smo analizirali s pomočjo plinske kromatografije (GC), plinske kromatografije sklopljene z masno spektrometrijo (GC-MS), plinske kromatografije sklopljene z masnim spektrometrom za analizo stabilnih izotopov lahkih elementov preko sežigne enote (GC-C-IRMS) in masnim spektrometrom z analizatorjem na čas preleta ionov (TOF) z elektrosprej ionizacijo (ESI Q-TOF MS in ESI Q-TOF MS/MS). Karbonizirani ostanki na posodah so AMS radiokarbonsko datirani. Rezultati kažejo, da so nekatere posode uporabljali za kuhanje mesa prežvekovalcev. V drugih posodah zasledimo ostanke kuhanja mešanega mesa prežvekovalcev in neprežvekovalcev ali ostanke kuhanja mešanice rastlinske in živalske hrane. Nekatere posode pa so bile uporabljene za predelavo mleka.
KEY WORDS - Eneolithic; pottery; organic residues; lipids; fatty acids; &3C values; TAG
Introduction
Organic residues survive in two principal forms on archaeological pottery: as charred deposits encrusted on ceramic vessels, and/or as absorbed residues preserved in low concentrations in highly degraded and complex matrices in vessel walls. Both are closely associated with the chaine operatoire of food storage, and food preparation, cooking and consumption.
In this article, we present direct chemical evidence for what we may recognise as preserved traces of
cooked, boiled, stored or processed food through li-pid characterisation using gas chromatography (GC), gas chromatography-mass spectrometry (GC-MS), gas chromatography-combustion-isotope ratio mass spectrometry (gC-C-IRMS) and soft ionisation electrospray mass spectrometric techniques ESI Q-TOF MS and ESI Q-TOF MS/MS.
The analyses were carried out on organic residues extracted from the Eneolithic vessels sampled from the Maharski prekop pottery assemblage. The analy-
ses focused on 20 potsherds (Tab. 1), covering well-defined groups of vessel size and related fabrics (see Mlekuž et al. this volume). Charred organic deposits on the surface of 22 vessels from the Maharski prekop assemblage were directly dated by the AMS radiocarbon method (Fig. 1). The dates demonstrate a distribution of probabilities, with periods of intensive occupation dating between 4400 and 4000 and between 3800 to 3550 calBC. While the charred deposits on pottery dating has been broadly accepted, Anders Fischer and Jan Heinemeier (2003; see also Craig et al. 2007) have noted that a number of dates obtained from vessel residues from prehistoric sites in northern Europe are too old with respect to the 'old' carbon effect, which was introduced into the residue through the processing of marine and freshwater organisms in vessels. However, we found no evidence that the pottery at Maharski prekop was used for freshwater fish, crayfish or mollusc processing.
The Maharski prekop site is located beside a palaeo-river channel in the Ižica flood plain. During the excavation in the 1970s, 2432 vertical wooden piles were recorded at the site. The arrangement of piles can be interpreted as the remains of nine houses with dimensions of around 10 x 3.5-4.5m arranged in parallel. The settlement is divided into two building phases. The pottery distribution within the 'cultural layer' is clustered in at least three distinct concentrations. However, we cannot confirm any direct relation between houses and pottery (see Mlekuž et al. this volume).
Palinological data indicate that the floodplain forest was mixed, comprising predominantly Quercus, Co-rylus, Fagus and Alnus, with some coniferous elements and open ground herbaceous taxa. In addition, the presence of cereal pollen is attested from at least 4000 calBC. A sharp decline in arboreal pollen and an expansion of herbaceous taxa, particularly cereals and Poaceae, and carbonised grains of Hordeum sp., support the notion of arable fields surrounded by forest (Culiberg, Sercelj 1991.252; Gardner 1999.130, 165, 168; Sercelj 1975.121; 1981-1982.104; see also Andrič 2007). The composition of the animal bone assemblage shows that domestic ruminant animals prevailed in local subsistence. The domestic package consists of sheep and goat (29.7%,), cattle (12.9%) and pig (5.1%). In addition, dog represents 12%. In the wild animal assemblage, red deer (18.1 %), roe deer (3.4%), boar (11.8%) and birds (3.0%) predominate (Drobne 1975a; 1975b).
Material and methods
We selected 20 pottery fragments for a pilot chemical study encompassing lipid distribution including fatty acids, stable isotope composition (bulk 513C or 515N, and 513C of individual fatty acids) and the di- and triacylglycerols (DAGs and TAGs) distribution of organic residues. Within the assemblage, three samples (MP85, MP158A and MP181) were obtained from charred organic residues from vessel surfaces (Tab. 1). The sherds were surface cleaned to remove any exogenous lipids. The sub-samples were then ground to a fine powder. First, the isoto-pic composition of carbon and nitrogen was determined on the bulk samples (Tab. 1). Each sample was weighed in duplicate into tin capsules and analysed using a Europa Scientific isotope ratio mass spectrometer with ANCA-SL preparation module for solid and liquid samples (PDZ Europa Ltd, Crewe, UK). Stable isotope results are presented as 513C or 515N values in per mil (%0) relative to the VPDB and AIR international standard, respectively.
Lipid analyses were performed using the extraction procedure described in detail by Richard Evershed et al. (1994). Two grammes of fine powder samples were transferred into a 50ml glass vial. Lipids were extracted with a mixture of chloroform and methanol (2:1 v/v), and 20|l of an internal standard (n-tetratriacontane, 1mg/ml in n-hexane) was added to determine lipid concentrations. The solvent was extracted by ultrasonication (2 x 15min) and evaporated to dryness under a gentle stream of nitrogen to obtain the total lipid extract (TLE). One portion of the TLE was derivatised using BSTFA (N, O-bis(trime-thylsilyl)trifluoroacetamide, 40|l; 70°C; 60min), evaporated to dryness and re-dissolved in n-hexane for GC and GC-MS analysis.
Further aliquots of the TLE were methylated using BF3/methanol to obtain fatty acid methyl esters (FAMEs) (14%, w/v; 100|l; Sigma Aldrich, Gilling-ham, UK; at 70°C for 1h). The methyl ester derivatives were extracted with hexane, and the solvent removed under nitrogen. FAMEs were re-dissolved in hexane for analysis by GC and GC-C-IRMS. GC-C-IRMS analyses were performed using an Isoprime GV system (Micromass, Manchester, UK).
The third TLE aliquot was used to identify the TAG content following the procedure described by Sigrid Mirabaud et al. (2007). The silica solid-phase extraction (SPE) cartridges (500mg, 3ml, Isolute, Biotage) were rinsed with 2ml of n-hexane, 2ml of
CH2CI2/CH3OH (2:1 v/v), and then 4ml of n-hexa-ne. The first fraction of the sample containing hydrocarbons was eluted with 1ml of hexane, while the second fraction, which contains the TAGs, was eluted with 1ml of CH2Ch. The fractions with TAGs were evaporated to dryness and then analysed. Mass measurements were run on a hybrid quadruple time of flight mass spectrometer (Q-TOF Premier) provided with an orthogonal Z-spray ESI interface (ESI-MS; Waters Micromass, Manchester, UK). The doping reagent of LiCl (Merck, Darmstadt, Germany) was added to the methanol sample solution for more efficient electrospray ionisation of long chain lipids. For ESI-MS/MS, the lithiated molecular ions [M+Li]+ were selected in the first quadrupole, accelerated to an energy of 3kV, dissociated by using 10 to 30eV collision energy and analysed with the TOF analyser.
Results and discussion
Bulk C and N isotope composition
The carbon and nitrogen isotope composition of the organic remains in pottery vessels may reflect the average 513C or 515N values of the degraded food during vessel use and burning. It may provide important information, particularly on whether the food was based primarily on animals or plants (C3 and C4). The determination of the isotopic composition of C and N was possible in only 15 samples. The average and standard deviation from potsherd samples are -30.7 ± 1.3%o and +1.0 ± 3.8%» for 513C and 515N, respectively (Tab. 1). These data fall in the range expected for degraded animal and plant tissues whose subsistence was based mainly on C3 plants.
Three charred organic residue samples from Maharski prekop (MP158A, MP85, and MP181) exhibit higher 515N values ranging from +6.2 to +9.8%. We suggest that these pots were used to process low-protein fat and/or high-fibre food (Spangenberg et al. 2006; Craig et al. 2007). However, the GC-MS analysis gives a more precise identi-
fication of food content. The C16:0/C18:0 fatty acid ratios of >2.0 and the presence of a homologous series of long chain n-alkanes from C16-C33 (odd-over-even carbon number predominance) indicate the use of plants and the presence of degraded vegetable oils in samples MP85 and MP181. In sample MP158A, on the other hand, the low C16:0/C18:0 ratio of 0.1, and the presence of cholesterol and stable isotope data with Ai3C value of -2.4% show the presence of ruminant adipose fat. These data match well with the results obtained for lipids in the bulk pottery sample MP158 from the same vessel, as we discuss below. The differentiating parameter between the two samples of a single vessel is the high 515N value of +9.8% determined in the charred organic residues on the inner surface of this pot (MP158A). It is worth pointing out that thermal degradation and microbial reworking of organic matter could significantly alter the 515N values, making this analysis less reliable, and should be used only as a general indicator prior to more specific chemical and molecular analysis such as lipid identification and stable isotope data.
Fig. 1. Radiocarbon dates of charred organic deposits on the surface of 22 vessels from Maharski prekop.
Lab. sample no.	Labcode AMS radiocarbon dating	Conventional radiocarbon age (BP)	S^Cbulk ±0.2 (%o)	S^Nbuik ±0.3 (%o)	Lipid concentration	Fatty acids	C16:o/C18:o	S'äC^o ±0.3 (%o)	±0.3 (%o)	Di- and triacylgricerols	Other constituents	Predominant commodity type
M Pi	Poz-48522	5280 + 40	-31-7	1-7	223	^■14:0, Cis-Ci^o, Ci6:o, C164, Ci8:i, Ci8:o, C2o:o, £22:0	1.8	-31.4	-31.4	n/a	odd carbon numbered higher alkanes	mixture
MP2	Poz-48659	475° ± 35	n/a	n/a	0.75	n/a	n/a	n/a	n/a	n/a	n/a	n/a
MP17	Poz-48514	4900 + 40	n/a	n/a	0.05	n/a	n/a	n/a	n/a	n/a	n/a	n/a
MP25	Poz-48510	5340 + 40	-29.9	3.8	434	Cio:o, C12:o, C1' 11 iS0-C,5;o, Ci6:o, C,8:1, C,8:0, C2o:o	2.0	-29.8	-29.9	D26-D34, T26-T34	C29H60	mixture
MP45	Poz-48512	5210 + 40	-30.4	-4.0	189	C,2:o, C,6:0f C,8:i, C,8:o	1.1	-29.6	-29.9	D30-D34, T26-T54	n/a	ruminant, possible goat milk
MP46	Poz-48513	4980 + 40	n/a	n/a	2.32	n/a	n/a	n/a	n/a	n/a	n/a	n/a
MP96	Poz-48516	5270 + 40	-29.7	2.2	129	Cio:o, Ci2:o, £13:0, Ci£:o, Ci8:o	10.7	-30.4	-30.3	n/a	wax, odd carbon numbered higher alkanes	plant
M Pi 00	P0Z-48502	547° + 35	-32.0	-0.4	3622	Ci2:o, Ci6:o, C,8:2, Ci8:i, Ci8:o, £20:0, £22:0	0.9	-28.3	-30.1	T4O—T 54	odd carbon numbered higher alkanes and alcohols, cholesterol	ruminant, cattle adipose fat
M Pi 21	P0Z-48503	5760 + 40	-32.0	3-5	110	Cg:of Cio:o, Ci6:o, Ci8:o, £20:0, £22:0	1.8	-30.7	-30.6	n/a	n/a	mixture
M Pi 23	Poz-48521	4790 + 35	-30.8	-0.2	870	^-■9:0, C10:o, C1g:0, Ci8:o, C22:o	0.2	-30.0	-29.8	D26-D36, T26-T54	odd carbon numbered higher alkanes	ruminant, possible goat milk
M Pi 44	P0Z-48504	5105 + 35	n/a	n/a	0.43	n/a	n/a	n/a	n/a	n/a	n/a	n/a
M Pi 51	Poz-48526	5000 + 40	-29.6	2.1	44-3	Cio:o, ^-444,, Ci6:of Ci8:o, £22:0	1.1	-30.4	-32.0	n/a	n/a	ruminant
M Pi 58			-29.8	-0.6	456	Cg:or Cio:o, C14/1, Ci6:of C,7:o, ^18:1, Ci8:o, C22:o	0.9	-29.1	-30.7	T4O—T 54	odd carbon numbere higher alkanes, cholesterol	ruminant, cattle adipose fat
MP158A	P0Z-48506	4860 + 40	-30.0	9-5	118	Ci6:o, Ci8:o, (-22:0	0.1	-28.0	-30.4	n/a	odd carbon numbered higher alkanes, cholesterol	ruminant
M Pi 72	P0Z-48508	4940 + 40	-28.9	"5-2	41.7	Ci6:o, Ci8:o, £22:0	1.2	-30.4	-31.1	n/a	n/a	ruminant
M Pi 74	Poz-48517	5310 + 40	-33.8	-2.3	57-3	C;41, C]o:o, t-l 441, i 5 0 - C j r;i4,, C,6:0, C,7:0, iS0-C,7:o, C,8:3, C]8:i, (-4 4 4,, C2o:o, (-22:0	2.4	-31-7	-31.2	n/a	odd carbon numbered higher alkanes, cholesterol	mixture
M Pi 77	Poz-48518	4700 + 40	n/a	n/a	0,28	n/a	n/a	n/a	n/a	n/a	n/a	n/a
MP211	P0Z-48660	4970 + 40	-30.3	2.6	9-5	Cg:o, Cio:o, (-444,, C144, Ci6:o, C,8:2, C,8:i, C,8:o, C22:o	1.4	-29.7	-29.5	n/a	C29H60	mixture
MP85			-26.7	9.8	22.8	(-14:0, C]£:o, Ci8:o	4.0	-32.5	-31.1	n/a	odd carbon numbered higher alkanes, C22:0 ketones	plant
M Pi 81			-25.2	6.2	21.7	^-■14:0, C1g:0, C]8:o	2.4	-32.2	-30.0	n/a	odd carbon numbered higher alkanes, C22:0 ketones	plant
Tab. 1. A summary of the analyses of the potsherds and organic residues from Maharski prekop.
Lipid identification in potsherds
The lipid residue analyses were performed with the GC and GC-MS methods. Lipid preservation was very good, with 75% of potsherds yielding an appreciable lipid concentration. Lipid concentrations of less than 5|gg-1 were considered unreliable and excluded from further interpretation (Tab. 1).
The well-preserved extractable lipids in 15 samples consisted mainly of free and bounded fatty acids identified by their methylesters (FAME) mass spectra. The most common distribution found in our samples was dominated by a high abundance of the Ci6:0 and Ci8:0 fatty acids derived from degraded animal fats (Fig. 2). The presence of degraded animal fats can be further recognised in the relative abundances of Ci6:0 and Ci8:0. In 11 samples (or 73%; numbers MP1, MP25, MP45, MP100, MP121, MP123, MP151, MP158 with MP158A, MP172, MP211) the C16:0/C18:0 ratios of fatty acids range between 0.1 and 2.0 (Fig. 3). They indicate ruminant adipose fat (Copley et al. 2005). In 2 samples (or 13%; numbers MP96 and
Fig. 3. The representative GC-MS total ion chroma-tograms of the fatty acids methylesters (FAMEs) with different C16.0 and C18.0 abundance extracted from the Maharski prekop pottery samples MP25, MP45 and MP100.
Fig. 2. The GC chromatogram displaying the trimethylsilyla-ted lipid extract from the Maharski prekop pottery sample MP158 that is characteristic of degraded animal fats.
MP85) the ratios are greater than 3.0 and consistent with degraded vegetable oils (Copley et al. 2005). The other two samples, MP174 and MP181, have C16:0/C18:0 ratios of 2.4, indicating the possible presence of dairy fats.
The C16:0/C18:0 ratios that range from 0.9 to 2.0 in 6 of the 11 potsherds extracts (MP1, MP25, MP45, MP100, MP158, MP211) and the ratio of >2.0 in MP174 are all consistent with degraded animal fats and clearly show evidence of fatty acid preservation. The main acids in these extracts are C120 C14:0, C150 iso-C15:0 C16:0, C17:0, iso-C17:0, C18:0, C18:1, C20:0 and C22:0. Other fatty acids were less frequently observed (C90 C10:0, C13:0). The isomers of Cm with two different double bond positions at 9 and 11 were found in all 7 pottery samples mentioned above. The presence of branched iso-fatty acids with 15 and 17 carbon atoms was observed in samples MP1, MP25 and MP174. The presence of these acids together with two double bonds positional isomers of Cm indicates ruminant animal fats that have been biosynthe-sized in the gut and rumen (Christie 1981; Dudd et al. 1999; Mottram et al. 1999; Regert 2011). The parallel biomarkers, i.e., the cholesterol and its derivatives, were detected in four samples (MP158 with MP158A [samples relate to the same vessel], MP100 and MP174), confirming the presence of degraded animal fats (Tab. 1 and Fig. 2). A single Cm isomer that reflects non-ruminant monogastric animal fat, such as porcine (Enser 1991; Evershed et al. 1997), was not detected in any of the samples from Mahar-ski prekop.
A homologous series of long chain n-alkanes from C16-C33 (odd-over-even carbon number predominance) that originate from epicuticular waxes of vascular plants (Bianchi 1995) was present in 9 samples (numbers MP1, MP96, MP100, MP123, MP158 with MP158A, MP174, MP85, MP181) (Tab. 1). The highest abundance of n-alkanes associated with the highest abundance of C16« in sample MP96 suggests the presence of plants as well. The correlation of long chain n-alkanes, the highest abundance of C16« and the highest C16VC18« ratio of 10.7 indicate that the vessel was used for preparing or storing plant foods. The appearance of both animal and plant bio-markers observed in the other 8 samples (numbers MP1, MP100, MP123, MP158 with MP158A, MP174, MP85 and MP181) suggests that the pots were used for mixed food processing and storage (Tab. 1).
Stable carbon isotope composition of fatty acids
The 513C values determined by the GC-C-IRMS method in fatty acids in modern edible oils and fats can be used to access the origin of lipids preserved in archaeological ceramics (see Copley et al. 2003; Spangenberg et al. 2006; Craig et al. 2007). The 513C values of the major fatty acids in oils from C3 plants are plotted near the 1:1 line, while modern dairy fats plot below the A13C = -3.3%o line (Evershed et
al. 2008; Copley et al. 2003; 2005; Dunne et al. 2012) (Fig. 4a). Seven samples of Maharski prekop plot near the 1:1 line, while the other five samples fall below the 1:1 line in a zone of ruminant adipose fats. No evidence of fish, porcine or ruminant dairy fat was found in the samples when observing the 513C values.
A more precise differentiation between non-ruminant adipose, ruminant adipose and ruminant dairy fats can be seen in the diagram where A13C values (513C18:o-513C16:o) are plotted against the 513C16:0 values (Fig. 4b). The comparison of the A13C values of modern reference animal fats with those of the Eneolithic pottery residues show that 42% of these plot in the range for ruminant adipose fats. A further 58% of the samples fall close to the limiting value between non-ruminant and ruminant meat (A13C = 0%). However, the later samples may not be assigned to meat mixture exclusively, but also to a mixture of plant-animal fats. The 513C values indicate that three pottery vessels (MP100, MP151, MP158 with MP158A) could contain sheep/goat and cattle fats. None of the 513C values determined in our samples plot below the A13C = -3.3% line, which was used as a criterion for the determination of dairy foods (Evershed et al. 2002; 2008; Copley et al.
Fig. 4. Plot showing: A the SUCma versus St3Ci6:0 values of modern reference animal fats and archaeological samples. B the difference in the S13C values of Cw.0 and Cum fatty acids (A13C) versus ftiCuM recovered from pottery extracts from Maharski prekop and modern reference fats (ft data from Craig et al. 2007; Gregg, Slater 2010; A data from Spangenberg et al. 2006).
Fig. 5. ESI Q-TOF MS mass distribution of the TAG fraction from the Maharski prekop pottery samples MP25, MP45, MP100 and MP123.
2003; 2005; Mukherjee et al. 2007; Dunne et al. 2012).
Distinguishing animal fats by ESI Q-TOF MS
DAG and TAGs are indicative lipids of degraded animal fats. Previous work has shown that the TAG distribution could help to differentiate between fats of ruminant and non-ruminant animals and ruminant dairy fats (Kimpe et al. 2002; Mirabaud et al. 2007; Regert 2011). We applied the ESI Q-TOF MS and ESI-MS/MS methods to obtain information about the structure of TAGs in the Maharski prekop pottery samples. A major advantage of this method is its sensitivity, which overcomes one of the major difficulties raised by the low occurrence of TAGs in complex archaeological samples. A similar approach including the nanoESI-MS method was successfully applied to a series of modern reference fats and archaeological samples (see Mirabaud et al. 2007; Garnier et al. 2007; Regert 2011). The TAG analysis of modern fats of ruminants (i.e., sheep, goat, and cattle), and non-ruminants (i.e., pig) has shown clear differences between these two groups. The ruminant adipose fats contain TAGs that range from T42 (for cattle) or T44 (for sheep) to T54 (Dudd et al. 1999; Mukherjee et al. 2007). The non-ruminant adipose fats also show a narrow distribution that range from
T44 to T54, but with low quantities of T44, T46 and T48. A broad distribution that ranges from T28 to T54 is recognised in dairy fats (Mirabaud et al. 2007; Regert 2011). It should be mentioned, however, that even if these distributions could partially resist degradation, they are modified in TAGs compositions to varying extents over time (Dudd et al. 1999; Mukherjee et al. 2007; Regert 2011).
The distribution of TAGs in five Maharski prekop samples shows that samples MP45 and MP123 have a large TAG distribution from T28 to T54. In the other three samples, a narrower distribution runs from T26 to T34 (sample MP25) or from T40 to T54 (samples MP100 and MP158) (Fig. 5). Along the saturated TAGs, two T50:i and T545 (sample MP45) four T50:i, T52:1, T54:3, and T54:2 (samples MP100 and MP158) unsaturated TAGs were detected. We compare these distributions with modern reference fats. We recognise the TAG distribution in samples MP45 and MP123 as indicators of dairy fat products (Fig. 5). The comparison of the TAG and DAG distributions in the identification of goat versus cow milk favours goat milk (Mirabaud et al. 2007). More precise identification could be obtained from fatty acid distribution in T44:0 using ESI-MS/MS fragmentation. This fragmentation was not possible in the Maharski pre-
kop samples due to the low amount and poor ionisation yield of T44:0.
Maharski prekop samples MP100 and MP158 show evidence of the presence of cattle adipose fats, since the TAG distribution starts at T40. This is in accordance with the Ai3C values for these particular samples (Fig. 4B). Some samples (MP25) show only low molecular TAGs and DAGs, which suggest that lipids are of animal origin, although more precise determination was not possible (Fig. 5).
On the other hand, the TAGs and DAGs distributions in the Maharski prekop samples (MP45 and MP123) do not correlate well with the A13C values. The values of -0.3 and 0.2%o are higher than expected for cow or even goat milk (Fig. 4B). One of the possible explanations for this could be milk processing, since it was observed that Ci6:0 and Ci8:0 fatty acids are enriched in 13C during heating, and thus the difference in the 513C values between the two acids becomes lower (Spangenberg et al. 2006). Further investigation is needed, therefore, to understand the influence of milk processing (i.e., warming and boiling) on the isotopic composition of fatty acids.
Conclusions
The results obtained from the lipid analysis of pottery from Maharski prekop indicate dietary trends involving animal and plant tissue. None of the total lipid extracts contained porcine or freshwater fish
adipose fats or ruminant dairy fats. A good correlation in the interpretation of ruminant adipose fats, based on lipid identification and stable isotope analysis, was obtained in pottery samples MP100, MP151, MP158 with MP158A and MP172. Furthermore, the TAG distributions in MP100 and MP158 samples correlate well with the presence of cattle adipose fats recognised in the discriminating criteria found in modern fats. In other samples, the presence of different mixed fats either from non-ruminant and ruminant meat or from a plant-animal fats mixture was observed. Nevertheless, the TAG distribution in two samples (MP45 and MP123) showed the residues of dairy products that probably derived from goat milk. These data do not match the outcome of the stable isotope analysis completely, since the A13C values in those samples are higher than expected for goat milk. One of the possible explanations for this distribution could be milk processing and dairy food production. Further investigation is needed to understand the inconsistencies that appear in the results obtained by different methods.
-ACKNOWLEDGEMENTS-
The research was undertaken as part of research projects J6-4085funded by the Slovenian Research Agency. We thank the Ljubljana City Museum and our colleague Irena Šinkovec for providing access to the Maharski prekop pottery assemblage. Special thanks go to Lucija Šoberl and Martine Regert for providing comments and suggestions.
References
Andrič M. 2007. Why were the Neolithic landscapes of the Bela krajina and Ljubljana Marshes regions of Slovenia so dissimilar? In M. Budja (ed.), 14th Neolithic Studies. Documenta Praehistorica 34:177-189.
Bianchi G. 1995. Plant waxes. In R. J. Hamilton (ed.), Waxes: Chemistry, Molecular Biology and Functions. The Oily Press, Dundee: 175-222.
Christie W. W. 1981. The effect of diet and other factors on the lipid composition of ruminant tissues and milk. In W. W. Christie (ed.), Lipid metabolism in ruminant animals. Pergamon Press, Oxford: 193-226.
Copley M. S., Berstan R., Dudd S. N., Straker V., Payne S. and Evershed R. P. 2005. Dairy in antiquity, I: evidence
from absorbed lipid residues dating to British Iron Age. Journal of Archaeological Science 32: 485-503.
Copley M. S., Berstan R., Dudd S. N., Dovherty G., Mukher-jee A. J., Straker V., Payne S. and Evershed R. P. 2003. Direct chemical evidence for widespread dairying in prehistoric Britain. Proceedings of the National Academy of Science 100:1524-1529.
Craig O. E., Forster M., Andersen S. H., Koch E., Crombe P., Milner N. J., Stern B., Bailey G. N. and Heron C. P. 2007. Molecular and isotopic demonstration of the processing of aquatic products in Northern European prehistoric pottery. Archaeometry 49:135-152.
Culiberg M., Sercelj A. 1991. Razlike v rezultatih raziskav mak-roskopskih rastlinskih ostankov s kolišč na Ljubljan-
skem barju in pelodnih analiz - dokaz človekovega vpliva na gozd (Differences between macro-remain and pollen analysis from pile dwellings - evidence for human impact on forest). Poročilo o raziskovanjupaleolitika, neolitika in eneolitika v Sloveniji 19:249-256.
Drobne K. 1975a. Živalski ostanki iz kolisca ob Maharskem prekopu iz let 1973 in 1974. Poročilo o raziskovanjupa-leolita, neolita in eneolita v Sloveniji 4:135-141.
1975b. Favna koliščarskih naselij na Ljubljanskem barju. Arheološki vestnik 24: 217-224.
Dudd S. N., Evershed R. P. and Gibson A. M. 1999. Evidence for varying patterns of exploitation of animal products in different prehistoric pottery traditions based on lipids preserved in surface and absorbed residues. Journal of Archaeological Science 26:1473-1482.
Dunne J., Evershed R. P., Salque M., Cramp L., Bruni S., Ryan K., Biagetti S. and Di Lernia S. 2012. First dairying in green Saharan Africa in the fifth millennium BC. Nature 486:390-394.
Enser M. 1991. Animal carcass fats and fish oils. In J. B Rossel, J. L. R. Pritchard (eds.), Analysis of oil seeds, fats and fatty foods. London, Elsevier: 329-394.
Evershed R. P., Arnot K. I., Eglinton G. and Charters S. 1994. Application of isotope ratio monitoring gas chroma-tography-mass spectrometry to the analysis of organic residues of archaeological origin. Analyst 119:909-914.
Evershed R. P., Mottram H. R., Dudd S. N., Charters S., Stott A.W., Lawrence G. J., Gibson A. M., Conner A., Blinkhorn P. W. and Reeves V. 1997. New criteria for the identification of animal fats preserved in archaeological pottery. Naturwissenschaften 84(9): 402-6.
Evershed R. P., Dudd S. N., Copley M. S., Berstan R., Stott A. W., Mottram H., Buckley S. A. and Crossman Z. 2002. Chemistry of archaeological animal fats. Accounts of Chemical Research 35(8): 660-8.
Evershed R. P., Payne S., Sherratt A. G., Copley M. S., Coolidge J., Urem-Kotsu D., Kotsakis K., Özdogan M., Öz-dogan A. E., Nieuwenhuyse O., Akkermans P. M. M. G., Bailey D., Andeescu R.-R., Campbell S., Farid S., Hodder I., Yalman N., Özbagaran M., Bigakci E., Garfinkel Y., Levy T. and Burton M. M. 2008. Earliest date for milk use in the Near East and southeastern Europe linked to cattle herding. Nature 455:528-31.
Fischer A., Heinemeier J. 2003. Freshwater reservoir effect in C-14 dates of food residues on pottery. Radiocarbon 45: 449-66.
Gardner A. 1999. The impact of Neolithic agriculture on the environment of South-East Europe. Unpublished PhD thesis. University of Cambridge. Cambridge.
Garnier N., Rolando C., Munk H0tje J. and Tokarski C. 2007. Analysis of archaeological triacylglycerols by high resolution nanoESI, FT-ICR MS and IRMPD MS/MS: application to 5th century BC - 4th century AD oil lamps from Olbia (Ukraine). International Journal of Mass Spectrometry 284: 47-56.
Gregg M. W., Slater G. F. 2010. A new method for extraction, isolation and transesterification of free fatty acids from archaeological pottery. Archaeometry 52:833-854.
Kimpe K., Jacobs P. A. and Waelkens M. 2002. Mass spec-trometric methods prove the use of beeswax and ruminant fats in late Roman cooking pots. Journal of Chroma-tography A 968:151-160.
Mirabaud S., Rolando C. and Regert M. 2007. Molecular criteria for discriminating adipose fats and milk from different species by nano-ESI MS and MS/MS of their tria-cylglycerols: application to archaeological remains. Analytical Chemistry 79: 6182-6192.
Mottram H. R., Dudd S. N., Lawrence G. J., Stott A. W. and Evershed R. P. 1999. New chromatographic, mass spectro-metric and stable isotope approaches to the classification of degraded animal fats preserved in archaeological pottery. Journal of Chromatography A 833:209-21.
Mukherjee A. J., Berstan R., Copley M. S., Gibson A. M. and Evershed R. P. 2007. Compound-specific stable carbon iso-topic detection of pig product processing in British late Neolithic pottery. Antiquity 81(313): 743-54.
Regert M. 2011. Analytical strategies for discriminating archeological fatty substances from animal origin. Mass Spectrometry Reviews 30:177-220.
Spangenberg J. E., Jacomet S. and Schibler J. 2006. Chemical analyses of organic residues in archaeological pottery from Arbon Bleiche 3, Switzerland - evidence for dairying in the late Neolithic. Journal of Archaeological Science 33:1-13.
Sercelj A. 1975. Analize makroskopskih in mikroskopskih rastlinskih ostankov s kolišča ob Maharskemu prekopu, izkopavanja 1973. in 1974. leta. Poročilo o raziskovanju paleolita, neolita in eneolita v Sloveniji 4:115-122.
1981-1982. Pomen botaničnih raziskav na koliščih Ljubljanskega barja (The significance of botanical investigations for the pile dwellings of Ljubljansko barje). Poročilo o raziskovanju paleolitika, neolitika in ene-olitika v Sloveniji 9-10:101-106.