Paula Kouki
Ten main and trace elements were analysed from soil samples taken from the dwelling site of Multavieru, and from comparison samples taken from undisturbed soil. Elemental analyses were carried out using FAAS for determining Zn and K and ICP-AES for Al, P, Fe, Mn, Ca, Mg, Cu and Sr. A principal components analysis was performed on the results. In addition to phosphorus, best indicator for past human settlement in Multavieru appeared to be the ratio of Ca to Mg. There were also anomalies of P, Ca, Zn, Sr and Mn in dwelling site soil.
Several main and trace elements are known to concentrate in soils of prehistoric dwelling sites as a result of human activity in the area. The enrichment of elements in soils is greatly dependent of several factors, the main three being climate, soil formation and the properties of soil (particle size, mineralogy etc.). Due to variation in these factors the enrichment of elements will vary from one site to another. The aim of my Master's Thesis was to find out which elements - if any - show anthropogenic enrichment in podzol soil and whether this enrichment could be connected to a specific human activity at the site. I also hoped to be able to suggest ways in which geochemical analysis could be used in future archaeological research.
The dwelling site chosen for this study is Multavieru in Polvijärvi, North Karelia, Finland. Department of Archaeology (University of Helsinki) excavated there for two months during 1996 and -97 and made a large scale phosphate mapping of the area. The dwelling site of Multavieru is situated on a low esker. Before the drainage of Lake Höytiäinen in 1858 (Vesajoki 1980: 3) it used to be on the shore of a long bay of the lake. (Lavento 1997b: 14.) The soil in Multavieru is fine sand. A well-formed podzol profile can be seen especially in those locations where human influence has been minimal.
Two shore terraces are discernible in Multavieru. On the upper terrace which is about 103 m asl there are remains of a dwelling site which dates most likely to the Mesolithic. The lower terrace, where the main site is situated, is 101 m asl. It has been inhabited possibly as early as the Early Comb Ceramic Period (5100-4100 BC). Concluding from the finds, the habitation seems to have been fairly continuous for the Neolithic and Early Metal Age. After that there has probably been a break in the activity at Multavieru, but there are again remains of settlement from historical times. The evidence for this is ceramics and remains of a building with a hearth. Small scale smelting of iron and tar-burning have also taken place during the 18th and 19th centuries. The main part of the dwelling site was destroyed in 1960's by the digging of a sand pit. (Lavento 1997a: 58.)
Samples for the analyses were taken in the course of the latter excavation. The sampling locations were not determined by a grid but chosen from different features and contexts in each excavation layer. For each sample ca. 250 g of soil was taken using a plastic scoop and placed in a small plastic bag. The samples were numbered and the location of the sample, sampling depth and colour of soil were recorded. For comparison samples were taken from outside the dwelling site area. The locations for comparison samples were chosen using the knowledge acquired from the phosphate mapping of 1996. The samples were taken from the area where the level of soil phophorus was lowest, it being most likely unaffected by human activity. Five test pits were dug and three to four samples were taken from each pit following the podzol profile.
Ten main and trace elements were chosen for analysis. These were Al, P, Fe, Ca, K, Mg, Mn, Cu, Zn and Sr. The analyses were carried out at the Geochemical Laboratory of the Department of Geology and Mineralogy (University of Helsinki) using FAAS (flame atomic absorption spectrometry) for detecting Zn and K and ICP-AES (inductively coupled plasma-atomic emission spectrometry) for the other elements. Ph.D. Antti Vuorinen designed a partial dissolution technique using hydrochloric acid in the pre-treatment of the samples. To dissolve organic complexes more effectively hydrogen peroxide was added to the reagent. This technique was considered sufficient as the elements enriched as a result of human activity would be likely to form complexes with the hydroxides of Fe and Mn and soil organic matter and absorb in clay minerals, which were dissolved by the treatment (Chao 1984: 101-111). The FAAS analyses were carried out by B.Sc. Tuija Vaahtojärvi and the ICP-AES analyses by Ph.D. Antti Vuorinen.
Other analyses were carried out as well: determination of soil pH and loss on ignition. Particle size distribution was determined by dry sieving of two samples taken for this purpose. In addition P was analysed using a colorimetric technique to compare the results of ICP-AES and the colorimetric procedure commonly used by archaeologists in Finland. Lastly the remaining sample was sieved with a 1,0 mm sieve and inspected with a petrographic microscope for microarchaeological remains of human occupation. It was hoped that the results of these analyses would give information of the sources of the element concentrations in the dwelling site soil.
The samples were divided into three groups and a statistical principal components analysis was performed on the results of the elemental analyses. In undisturbed podzol soil the behaviour of all the elements was remarkably similar and no pronounced grouping of elements appeared in the principal components analysis. The samples from the upper terrace resebled those of the natural podzol except for enrichment of Mn, Zn and Mg in the B-horizon. In the dwelling site samples, however, three distinct groups of elements formed. The main factors in the first principal component from the dwelling site samples were Cu, Zn, Sr, Mn and Ca. In the second principal component the most important elements were Al, P and Fe, while the influence of K and Mg was strongly negative, and the third principal component was mainly characterised by K and Mg.
pH-values and sampling depth were plotted and correlated against the principal components. It could be seen from the correlation matrices and scatter plots that in natural podzol soil pH and sampling depth correlated with the distribution of K, Mg and Ca (but surprisingly enough, not with Sr). These elements are easily soluble in acid environments and their amount increases deeper in the ground where the pH is higher. On the upper terrace similar correlation with pH and depth was seen for K, Sr and Ca, but not for Mg. For some as yet unexplained reason Mg is enriched in the upper terrace soil despite it being more easily soluble than any other of the aforementioned elements.
On the basis of the results from the statistical analyses Cu, Zn, Sr, Mn and Ca were considered to be the elements most likely to give information of past human activity at the Multavieru dwelling site (in addition to P). However, a closer examination of the results revealed that the amounts of Cu in dwelling site soil are mostly similar to those in the comparison samples and there are only a few, rather weak anomalies of Cu.
Ca appeared to be the most interesting element in Multavieru. In natural podzol the curves for Ca and Mg are alike (fig. 1), the amount of both elements increasing with depth and being greatest in the C-horizon. This is most probably due to the chemical similarity of these elements (Kabata-Pendias & Pendias 1985: 20-21, table 9). In dwelling site soil, however, the amount of Ca is highest in the B-horizon and decreases somewhat in the C-horizon, whereas the behaviour of Mg is similar to that in natural soil (fig. 2). There is also considerable enrichment of Ca in several samples from the B-horizon of dwelling site soil, especially from the Neolithic layers. Plenty of fragments of burnt bone were found in these layers in the course of the excavation. It seems likely that the enrichment of Ca at the dwelling site is a result of decomposition of bones. It can be supposed that in addition to the burned bone material found, there has been considerable amount of unburned bones, which are archaeologically "invisible" since they have completely dissolved in acid soil.
The anomalies of Sr seem to be associated to the enrichment of Ca. Sr is chemically similar to Ca and it is known to associate with Ca in soils (Kabata-Pendias & Pendias 1985: 94). However, in the comparison samples there is no close connection between Sr and Ca. It is possible that Sr in dwelling site soil is derived from the same source as Ca. There are small amounts of Sr in bones for example.
There are also anomalies of Zn in dwelling site soil. These occur with anomalies of other elements, usually with Ca and P. It may here be of interest that most of the burned bones found were fish bones, which contain comparatively large amounts of Zn. It would be tempting to associate the enrichment of Zn, Ca and P in the same samples to the fish bones, but considering the nature of the Multavieru site this might be a hasty conclusion because it is simply impossible to discern what effect the different periods of settlement have had on the element content of soil.
Sometimes the anomalies of Zn are also associated with those of Mn. At the dwelling site the anomalies of Mn seem to occur with black, sooty soil. It must be noted, however, that the amount of Mn can vary greatly in natural soil as well. As the hydroxides of Mn are very effective in adsorbing heavy metals, enrichment of Mn in natural soil can also lead to enrichment of Cu, Zn etc (Chao 1984: 103). This may be the cause of Zn anomalies in some upper terrace samples in which Mn is considerably enriched. In one of the test pits there is also an anomaly of Mn, but no exceptionally high amounts of Cu or Zn are observed there.
The distribution of K and Mg in dwelling site soil is similar to that in natural soil. Normally the amounts of both elements increase downwards in soil. In one profile in Multavieru there was, however, a distinct peak of K at the depth of ca. 40 cm, combined with a smaller increase of Mg. At the same depth the amounts of Al, P and Fe decreased to increase again deeper in the ground (fig. 3). Together these observations indicate that there has been an old topsoil which was buried sometime during the break in human settlement in Multavieru. This assumption is supported by the results of microscopical examination of the samples: no microarchaeological remains were found in this layer.
Three samples were taken from a Neolithic hearth, where the soil was very sooty and felt greasy to touch. There was also a considerable amount of burned bones in the hearth. These samples showed very high anomalies of almost all the elements analysed, excepting P, Al and Fe, the amounts of which agreed with natural podzolisation process. The anomalies are likely to be derived from the material that has been burned in or otherwise has ended up to the hearth and/or they may have concentrated later to the hearth from surrounding soil.
Reliable conclusions cannot be drawn on the basis of the analyses of one dwelling site only, and the following should therefore be considered merely as tentative suggestions for the possibilities of elemental analysis in archaeology. More research is needed to confirm these suggestions - or to discard them.
On the basis of the Multavieru analyses it seems that, in addition to P, the best indicator for past human settlement is the ratio of Ca to Mg in soil. Considerable enrichment of Ca was observed especially in the cultural layer formed during the Neolithic hunter-gatherer settlement. It might be possible to use the Ca/Mg -ratio for recognizing prehistoric settlement when the results of phosphorus analysis are conflicting or when lack of finds from test pits makes it difficult to decide whether a possible site really is a dwelling site.
Due to the multi-period nature of the Multavieru site it is not possible to reliably connect the enrichment of any element to a specific human activity. However, it appears that different contexts and features can be recognized by their element concentrations. The locations containing burned bones were characterised by very high amounts of P and Ca, and in many cases also Zn and Sr were enriched. Enrichment of Ca, Zn and Mn with no exceptionally high amounts of P seems to be connected with sooty soil.
The results from Multavieru seem also to indicate that elemental analysis could be useful at dwelling sites with a succession of settlement phases and complicated stratigraphy. It could be used for separating different layers and recognizing the layers of human occupation. It might also be of use in ascertaining that certain layers in separate excavation areas belong to the same context.
Chao T.T. 1984: Use of Partial Dissolution Techniques in Geochemical
Exploration. Journal of Geochemical Exploration 20
pp. 101-135.
Kabata-Pendias Alina & Pendias Henry 1985:
Trace Elements in Soils and Plants. 3rd printing.
Boca-Raton, FL CRC Press.
Lavento Mika 1997a: Dwelling Sites of Sarsa-Tomitsa Ceramics in
Finland and Karelian Isthmus. Part 2. Unpublished Licentiate's
Dissertation.
Lavento Mika 1997b: Polvijärvi Kinahmo Multavieru. Kertomus v:n 1996
kaivauksesta. Unpublished report of excavation.
Vesajoki Heikki 1980: Pre- and Post-Drainage Development of the Shore
Morphology and Stratigraphy of Lake Höytiäinen, Eastern Finland.
Joensuun korkeakoulun julkaisuja - Publications of the University
of Joensuu sarja BII 13.
