The Project successfully completed a five-week summer season, running from 15th July to 17th August 2013. The principal objectives of the 2013 seasons were defined before the season as follows:
The British excavations: the aim of excavating a pit and a linear feature in close proximity to each other, as well as to a group of houses selected for test-pit investigation, led to the selection of an area in the South West part of the mega-site under permanent grassland.
In this area, five houses (in Group 1) were subject to test-pitting; a single trench (Sonda 2) was laid out across a curved linear feature which appeared on the geophysical plot as a ditch delimiting part of the outer circuit of houses; and a larger trench (Sonda 1) was laid out over a magnetic anomaly consistent with a pit. The five test-pits are reported on later.
Under John Chapman’s supervision, a trench 3m in length and 1m in width was laid out over the top of the suspected ditch. The trench was excavated to a depth of 1m, with the A horizon of the chernozem recognisable between the surface and 0.45m depth. Two possible cut features were noted – one at either end of the North section – but these features were inconclusive. The trench was extended both to the East (by 0.65m) and to the West (by 1m), making a total size of 4.65m x 1m. There were no secure signs of ditch cuts in either extension.
Occasional Trypillia sherds were found in the chernozem B horizon, which was also characterised by intensive precipitation of white carbonate throughout the B horizon, to a depth of 1m (this carbonate precipitation is a purely natural pedological process frequently observed in chernozems). The extended trench was excavated to within 10cm of the chernozem C horizon (aka the ‘dirty loess’), with no further possible cultural features encountered.
In summary, the result of the excavation of Sonda 2 was the definition of a natural chernozem profile, with all three horizons, to a depth of 3.00m. We have to conclude that there was no correspondence between the buried features and the magnetic anomaly consistent with the line of a sub-surface ditch.
A trench (termed Sonda 1), 6.70 - 6.95m East – West x 3.97 - 4.07 m North – South was set over a large geomagnetic anomaly. In the South wall of the trench, an identical profile to that found in Sonda 2 was encountered: an A horizon to 0.45m depth, with a B horizon dominated by carbonate in-washing to a depth of 1m.
The first 40 cm from the surface were excavated over the entire area in two 20-cm spits. Although there were sherds in both layers, clear concentrations of archaeological materials and pit boundaries were not identifiable. This imposed a change of excavation strategy, with subdivision of the trench into a North and a South sector. In the course of the excavations, it became obvious that it was not possible to identify clear-cut pit boundaries in terms of differences in colour or soil consistency. A working hypothesis for this situation based on observations during the excavations and discussions with Ukrainian colleagues is that it is related to the chernozem formation, on the one hand, together with more than one episode of deposition, on the other hand. These have led to 'blurring' of the surface/s from which the pit was dug and the only way of defining the pit was the concentration/s of sherds and finds. Following these concentrations proved to be a valid strategy for excavating deeper, since spreads of sherds (sometimes very large) were found in a light yellowish/beige sediment believed to be the sterile matrix/the bedrock.
A combination of trowelling and spit digging was employed in order to reconcile speed with precision of recovery. Flotation samples were taken at regular intervals. The study of the mode of deposition is very much in its early stages but a general pattern throughout the pit matrix is the repeated occurrences of horizontal spread of sherds, bones, tools, figurines, ground stone and burnt and unburnt daub. Typical examples are shown below.
While sterile soil/bedrock was reached in the Eastern part of the pit, a small test pit in the south-western part of the pit suggested a minimum additional depth of 40 cm of pit-fill there. Therefore it was decided that the investigation should be completed during the next field season and the trench was back-filled after appropriate preparation.
A total of 6,677 sherds and 20 figurines was found in the 2013 excavations.
The Ukrainian side laid out an ambitiously large trench with the aim of investigating three contrasting types of features: (1) two adjoining burnt houses close to the mega-structure excavated in 2012; (2) large pits associated with each of the two burnt houses; and (3) parts of the so-called ‘linear pits’ surrounding House A 19 on three sides.
Burnt houses: one complete burnt house was excavated (House B17), as well as part of a second house (House B18). The shapes of the two-storey houses corresponded to the shape of the anomalies on the geophysical plot.
The dimensions of House B17 were 24 m in length by 8 to 12 m in width. Most of the area of the house was covered in destruction daub to a depth of over 0.30m (viz., the classic ‘ploschadka’). Several abrupt changes in the slope of the burnt daub suggested the existence of pits or hollows under the daub layer; however, this did not turn out to be the case. Instead, the form of burnt daub was probably connected to the form of house destruction.
Interior details from House B-17 show incised and red-painted decoration; in particular, near the edge of the house, plaster fragments were found to have painted decoration Under the daub were two raised areas decorated with incised ornament, comparable to those on the ‘altars’ in the mega-structure. Several sherd scatters in House B-17 corresponded to groups of once-complete pots; one such vessel had painted decoration in the shape of a dog.
Dr. Manuel Arroyo-Kalin excavated a test section through the ploschadka and into what was thought to be under the floor level in order to provide a soil micromorphological characterization of the house floor and what had preceded its construction.
Dr. Arroyo-Kalin sampled a section of the B17 pit for soil micro-morphological analysis.
The pit to the North of House B18 held fewer finds but was even larger than the B17 pit, although only an 8m-long section was excavated to a depth of 2.5m. The upper fill, to 1,2m depth, was a chernozem; under this layer, the cultural layer shared the same properties as in the B-17 Pit, with a total of four anthropomorphic figurines.
It is clear that the initial; use of both pits was to extract clay for house-construction. Both pits were partially filled-in at the time of dwelling. The pits were still visible as negative features at the time of the end of the settlement: the upper fill consisted of a lower layer of soil mixed with cultural material, and an upper layer formed by chernozem. The linear-pits were dug on three sides of House B17 but were much more shallow than the B17 or B18 pits. The primary use of all of these pits could have been to produce soil for mixing with clay in house construction, as in the LBK system of digging pits close to houses.
These pits were much more shallow than the pits at the short end of the houses – usually no more than 30cm depth. In the linear pit near House B-17 were found sherds, animal bones and fragments of three anthropomorphic figurines. The further the linear pit was laid out from the house, the fewer the finds that were discovered. By the time of the house-burning, and unlike the larger pits, the linear pits had filled up to the general surface level.
All of the pits and linear-pits under excavation in 2013 were located through the occurrence of pit-like anomalies in the 2009 geophysical prospection. While the location of the pit-like anomalies was confirmed by excavation in each case (as were the locations of each burnt house), the anomalies were far smaller than the eventual size of the pits. This is unusual, since the apparent extent of a geomagnetic anomaly is typically larger than the material causing the anomaly. It must be assumed in this case that the contrast between the magnetic susceptibility of the material at the edges of the pits and that of the subsoil was insufficient to have generated anomalies. Perhaps it is the concentration of fired ceramic towards the centre of each pit which gives rise to the anomaly, rather than the backfilled sediments. The single anomaly which was not confirmed through excavation was the linear anomaly next to Sonda 1; neither ditch cuts nor ditch fill were encountered in the test pit.
The results of the 2013 environmental processing are best summarised in four sections:
Methodology: The 2013 season was very productive, and 98 soil samples, totalling c.1200 litres, were processed by a team of 2 over c. 4 weeks. All processed material was rapidly assessed and characterised on-site. We used bucket flotation rather than the flotation tank, which is ideal for large volumes (e.g. 40 litre +) and charcoal rich samples, but inefficient for smaller volumes (c.10-20 litre), or samples with low total charcoal abundance. Bucket flotation also recovered a wider range of macro-fossils, particularly mollusc shell and low buoyancy charcoal. The simple equipment needed for bucket flotation means it can be scaled up quite easily to allow more people to be involved with processing/sorting samples, including Nebelivka locals. A larger processing team could allow for more intensive environmental sampling programs, as-well-as greater recovery of small bone fragments, flint debitage or other small artefacts easily missed in excavation.
Charcoal for AMS C14 dating: Of key importance to the project is the recovery of datable material, and if it were not for the very interesting questions raised, the generally very low quantities of charcoal recovered might be disappointing. Charcoal was nearly absent in contexts directly associated with house structures, most notably in relation to the daub destruction layers. If present at all, charcoal was nearly always extremely fragmented (with obvious taphonomic problems for precision AMS C14 dating).
Occasionally daub found in the destruction layers has been burnt in low oxygen conditions, giving a grey-to-black colour. Rarely fine carbonised organic temper can be observed adhering to and 'interwoven' with the clay fabric. Three recovered examples are highly suitable for AMS C14 dating. Detailed investigation of the temper (non-wood/grass?) is possible in at least one. Four more have very low amounts of charcoal, and are probably not viable for dating. Successful dates from daub temper would relate very precisely to the creation and mixing of a particular batch of daub, although not necessarily to the initial construction of a building.
Archaeobotany: Perplexing as the low quantities of wood charcoal are, the apparent absence of any carbonised cereal grains or cereal chaff is remarkable. I am highly confident in the particular recovery methods used; the results compare with the recovery of only 1 ½ cereal grains (and no chaff) in 2009. Additionally, during examination of many daub fragments, I did not see any impression that could be clearly identified as grain, straw fragments, or other cereal chaff. Structures in the daub looked quite narrow, flat, and grass-like. Also, one clear example of a deciduous leaf impression (oak?) was spotted by the Ukrainian excavation team. Further investigation of daub temper would be useful. A provocative research question for the site would seem to be “Was there ever large quantities of grain or grain processing on the site, and if so, where did all the waste products go?”.
No other carbonised seeds were recovered, and only a single fragment of nutshell (hazel?) was found. Identification of wood charcoal species is possible in some samples, even if just to generic levels, and this could help explore the issues of woodland cover and prehistoric landscape management outlined below. Additionally, we did not have time to process all the samples from the large pits – these are likely to produce useful data when more complete.
Palaeoecology: Using bucket flotation, we recovered small-to-moderate assemblages of snail shell, across a range of samples.
This is a new strand of data to the project, and hopefully will complement regional and local pollen diagrams, archaeobotanical data, and micromorphology/ geomorphology. In 2013, test-pits were dug to examine deep profiles and the potential of long-term site-specific paleoecological data. The first test-pit recovered samples from an area away from known archaeology.Vallonia excentrica seems to dominate nearly every sample, and there are notable absences of many snail species indicative of climax Holocene forest. Indeed snail diversity seems slightly higher earlier in the Holocene, with the assemblages potentially indicative of a mix of established grassland and colonising woodland. By the time of the mega-site, and probably significantly earlier, snail faunas became less diverse, and indicate uniform grassland over a significant local area.
This could imply significant human landscape management prior to the mega-site, with perhaps a role for large animal herds. The Holocene palaeo-ecology of site can be explored further, using snails as one ecological proxy. Snail shell could also be a source for Carbon/Oxygen isotope curves, and maybe even for high precision Uranium-Thorium dating.
Closer investigation of macro-fossils from archaeological deposits is also likely to produce good data. One interesting result has been the occurrence of small numbers of Lymnea / Succinea shells, and rare fragments of freshwater mussel. These marsh/aquatic species are particularly found in shallow reed beds. Their occurrence in deposits associated with house structures could suggest the transport and domestic use of reeds. Given the lack of clear wood impressions or wood charcoal, and the abundance of thin-flat daub impressions, it might be worth exploring whether reeds could also form part of the house structure itself.
The availability of two large stubble fields in the North Eastern and North Western parts of the mega-site enabled the completion of a further 130 ha of geophysical prospection in 2013 by a team of five Durham geophysicists – Natalie Swann, Richie Villis, Trish Edwards, Ashley Hayes and Andy Blair. This brings the total area investigated in three seasons to ca. 200 ha, or 80% of the mega-site
The new geophysical investigations confirmed the principal structural features of the mega-site plan – the outer and inner circuits of burnt houses, with a large, mostly empty, area between the circuits; a number of radial streets laid out at right angles to, and inwards from, the inner circuit; a scattering of house-shaped anomalies outside the outer circuit; an empty area at the centre of the site (in the doughnut model of the mega-sites, the ‘hole in the doughnut’); and the continued recognition of pit-like anomalies and anomalies likely to represent unburnt or partially-burnt houses.
Several features were, however, surprising in the 2013 results:
(a) The large number of radial streets, which now surpasses 50 streets
(b) The occurrence of a linear anomaly concentric to the outer circuit, that can be interpreted a ditch running for 250m around the North East and North West parts of the mega-site
(c) Several more examples of rectilinear clusters of houses, generally comparable to the find of ‘Nebelivka Square’ in 2012
(d) The discovery of further very large structures, often in liminal places, of the scale of the mega-structure excavated in 2012. This includes an oval structure which appears to resemble a prehistoric megalith rather than the normally rectangular Trypillia house
(e) The variability in spacing of house groups
(f) Following the information from Taljanky of the discovery of very strong, small, circular magnetic anomalies that may represent pottery kilns, the discovery of two such strong anomalies in an area of Nebelivka with houses close by.
In summary, geophysical prospection in summer 2013 covered an additional 130 ha, with five geophysicists using three Bartington gradiometers over the 5-week season. The great progress made in the 2013 geophysical prospection means that we are coming close to a situation where we can make calculations of the total number of structures in the mega-site. However, this total represents a maximum number of structures, and not an estimate of the number of houses in coeval occupation. The modelling of AMS dates from different house groupings requires the collection of a wide range of samples for dating.
In 2012, the preferred strategy to advance towards an internal chronology of the Nebelivka mega-site was coring into the burnt daub of a large number of Trypillia houses, located by geophysical prospection, for the recovery of burnt organics for AMS dating. Technical issues with this approach, however, necessitated a revised approach, in which test-pits were excavated in the centre of burnt or unburnt houses located by geophysical survey.
In accordance with the AHRC application, a sampling strategy was worked out by Andrew Millard, Tony Wilkinson and John Chapman to determine the best way to model AMS dates for different sectors of the inner and outer circuits, as well as a selection of radial streets. This consisted of a total of 23 house groups from the 2009 and 2012 areas, together with a number of spaces in the sampling programme to cover geophysical findings in the 2013 season. However, on arrival at Nebelivka, it became clear that heavy crop cover over an estimated 70 ha vitiated the sampling strategy. The basic principles of what should be sampled were maintained but a different suite of house groups were selected, based on ready availability.
A second consideration was the size of the test-pit and the time required to excavate such test-pits. Under Ukrainian law, the maximum size of a test-pit (Ukrainian ‘shurf’) is 5 sq.m. The co-directors agreed to try out a standard 4m x 1m test-pit in the first group of houses.
The results of the test-pitting programme were impressive. Thirty-nine test-pits were excavated – 5 at 4 x 1m and 34 at 2 x 1m, with extensions in two cases. The general stratigraphy of the test-pits showed four contexts:
(a) upper part of Chernozem A horizon = topsoil
(b) ploshchadka level of burnt daub – sometimes sub-divided into earlier and later destruction daub levels;
(c) floor deposits with(out) built features (pits, ovens, ‘platforms’, etc.)
(d) lower part of Chernozem B horizon (pre-house)
In many test-pits, it was difficult to identify floor deposits, especially since fired clay floors and stamped earth floors did not appear to be a common cultural practice. This was especially so for the few unburnt houses under investigation. For this reason, Dr. Arroyo-Kalin sampled three sequences going from the ploshchadka to the pre-floor deposits for soil micromorphological analysis.
Each test-pit was excavated to the base of the pre-house level, followed by 0.50m of further excavation well into the Chernozem B or sometimes the C horizon, to ensure that there were no further pre-house features.
As may be expected, the architectural remains in the 39 investigated houses were markedly different, with variations in the quantity of daub and its spatial distribution, the proportion of vitrified to non-vitrified daub and the nature of inclusions placed in the daub. The Project realised that the test-pit approach provided an excellent sample of architectural remains across a wider number of houses than is usually excavated in a single season at a Trypillia site; a systematic collection of daub samples provided our building specialist, Dr. Nataliya Shevchenko, with an unparalleled set of data with which to investigate the clay sources used for houses, the construction techniques and the destruction techniques.
The number of samples per test-pit varied from zero (three cases, even after the excavation of extensions) to 22 (Group 13/House 3). In each of the five houses with a single AMS sample, the sample was of animal bone. There are no test-pits with only charcoal or shell or burnt daub samples. This means that, if the nitrogen test shows the presence of collagen in the animal bone samples, there is a good probability of dating 36 houses this year at the Oxford Facility, with the collaboration of Professor Tom Higham.
The change in the strategy for recovery of AMS samples from the mega-site houses meant that the mechanical coring in summer 2013 was less focussed on Trypillia houses than in 2012. The coring team of Joe Roe, Sean Hamer, Iunn Jenn Ong and Max Ratcliffe took two cores to investigate a linear anomaly (? ditch) found during geophysical prospection. However, activities focussed on two coring transects across the mega-site, in both North – South and East – West directions. The linking of the core points to a GPS record means that the two main coring profiles can be created in GIS, with additional detail from individual coring points (houses in 2012; test-pits in 2013).
The two on-site cores were placed to investigate the linear anomaly concentric to the outer circuit of houses – with an initial interpretation being a ditch. The first core was placed in the North East part of the linear anomaly and reached a depth of 5.50m, without hitting any obvious ditch fill. Instead, there were two principal deposits in the core: a lower reddish silty clay deposit 1.79m1 in width (4.29m – 2.50m) and an upper off-white silty clay deposit 1.80m in width (2.50m – 0.30m). Informal testing of these clays showed that both were suitable for pottery-making. It is currently hard to explain how such thick clay deposits came to be present in a feature that may have been a ditch.
The second core through the linear anomaly was placed in the North West part of the mega-site. At the base of the 4.50m-deep core, a buried chernozem C horizon had developed over 1.10m (4.50 – 3.40m), with a 1.40m-thick deposit of alluvial clay above the first C horizon (3.40 – 2m). Above the alluvial clay, a typical chernozem sequence developed with an A, a B and a C horizon. Intriguingly, the contents of both cores into the so-called ‘ditch’ differed markedly from each other, as did the types of clay found in the two cores. This means that a priority for excavation in summer 2014 is a section cut across at least one part of the linear anomaly.
In a five-day research visit to Nebelivka 2013, Dr. Manuel Arroyo-Kalin was introduced to the South Ukrainian landscape, the Nebelivka micro-region and the mega-site itself. Owing to limitations of time, he restricted his soil sampling to locations on the mega-site itself, in response to four research questions:
(a) The possibility of Middle – Late Holocene re-deposition of older loess contributing to post-Trypillia soil build-up: column and bulk samples were taken from the soil pit near the centre of the site
(b) The composition of Trypillia pit-fills: column and bulk samples were taken from two pits – what transpired to be the upper and middle phases of the large pit near House B17; the upper and middle phases of Pit 1, in the South West part of the mega-site
(c) The pre-house deposits, house construction and burning phases of Trypillia houses: samples were taken from the only completely excavated house in 2013 – House B17, as well as from five houses investigated by test-pitting. One unburnt house and five burnt houses were investigated through undisturbed block and bulk samples particular attention will be paid to the construction of the floors, which seemed in most cases rather poorly defined.
(d) Identification of buried soil surfaces in the kurgan at the Northern end of the mega-site the barrow at the Northern end of the mega-site had been looted in the 1970s or earlier, leaving a large looters’ pit in the centre of the mound. A small test-pit was excavated from the base of the looters’ pit, with a 3.5 m-deep section which was cleaned up in preparation for interpretation and sampling.
Four interesting results came from the test-pit:
(a) signs of a burial pit showed up in the section; (b) there were signs of buried A, AB and B horizons in the section, despite multiple krotovina intrusions; (c) small quantities of comminuted daub and very rare heavily fragmented Tripillia pottery was found in the buried A horizon, suggesting, alternatively, that either part of the barrow mound was created from soil over the mega-site cultural layer or that the observed buried A horizon is redeposited sediment containing abundant settlement related debris; and (d) the height of the buried A horizon was over 0.25m above the current level of the field near the barrow, where the Trypillia cultural layer was buried by an overburden of at least 0.25m. This suggests active erosion processes in the post-Trypillia period, which led to the removal of at least half a metre of chernozem A horizon. These results may be corroborated in the sediment cores taken by Dr. Bruce Albert in the alluvial zones adjoining the mega-site.
Now that the soil micro-morphological samples have been successfully brought to the UK, Dr. Arroyo-Kalin will make a preliminary assessment of the database with a view to developing a funding application for the further investigation of the Nebelivka samples.
Intensive, systematic fieldwalking of a further 18 sq. km in the 5km radius of the mega-site was led by Mr. Marco Nebbia, Ms. Kirrily White and Mr. Dan Miller. A total of 32 fields were walked at 20m spacing, with collection of surface material in 50-m lengths. The GPS locations of all sites were recorded, together with GPS spot points for significant off-site discard.
The results indicate the absence of Trypillia scatters or even off-site discard, in comparison with the presence of one LBA / EIA site (Site 1 / Field 25) with Greco-Roman import sherds and one off-site discard of this period (one handle fragment in Field 42). There were four Chernyakovska culture sites (Sites 1, 3 – 4: Fields 25, 51 and 55), on one of which was also found a thin scatter of AD 12th Medieval sherds. Recent finds were found in small quantities on many fields but in greater densities at Sites 1, 2 and 4 (Fields 25, 26 and 51) and as off-site scatters in Fields 42 and 45.
The identification of barrows in the fieldwalked area shows a relatively high density of what are assumed to be post-Trypillia mounds. On account of their small size, these barrows can most probably dated to the EBA rather than the Scythian periods. One other possible monument was the possible earthwork located in Field 44.
In terms of the Trypillia focus on the settlement pattern within a 5-km radius of the mega-site, this year’s fieldwalking results revealed even less Trypillia discard than in 2009 and 2012, viz., four small scatters and some off-site material in 2009 and one single large off-site painted sherd in 2012. This means that, so far, no Trypillia sites coeval with the mega-site have yet been found; moreover, very little Trypillia off-site discard has been found. But the other absences are also interesting, viz., the absence of Trypillia sites either earlier or later than the mega-site, which is securely dated to Phase BII. There is a possibility that some of the chipped stone finds made in Fields 30, 31 and 54 as part of off-site scatters may be dated to the Trypillia period.
In the post-Trypillia period, a small number of mortuary barrows has been located, suggesting an alternative approach to monumentality than was used in the Trypillia period. This switch from monumental rectangular buildings – especially if the Trypillia houses were sometimes two-storey – to monumental circular mortuary barrows suggests a major change in cultural values and the use of the landscape. No evidence of the settlements relating to the phase of barrow construction has yet been discovered.
For later periods, fewer LBA / EIA sites (n = 2) have been found than Chernyakovska group sites (n = 5), suggesting an increase in settlement numbers in the period contemporary with the Roman period in Western Europe. However, there are still gaps in the settlement record, with LIA and Early Medieval (AD 5th – 11th) sites hard to locate.
The total area that has now been fieldwalked amounts to 28.5 sq. km – or some 40% of the total area within a 5km radius of the mega-site, excluding the area of the modern village - for 2009 = 55ha; for 2012 = 10 sq. km; for 2013 = 18 sq. km. Thus, the Project still has the target of fieldwalking additional fields within the 5-km radius as and when fields become available.
Satellite imagery and fieldwalking
The fieldwalking season was structured around the results of a detailed remote sensing analysis based on high-resolution satellite imagery (WorldView-2). This is one of the best satellite dataset available at the moment, both in terms of spatial and spectral resolution: the data acquired include a panchromatic image, with a spatial resolution of 0.50 m, and an 8–band image with a spatial resolution of 2.00 m. After a pan-sharpening process, it was possible to “raise” the spatial resolution of the multispectral image to 0.50 m and then performing several image processing techniques in order to enhance the visibility of features of interest.
The photo-interpretation included the entire area covered by the image (100 sq. km around the mega-site). The set of anomalies mapped comprised three main classes of features with different shapes and spectral signatures. Therefore beside the plan of walking every available field within 5 km from the mega-site, the field-survey was focused on the ground-truthing of such features in order to check and refine the remote sensing interpretation.
The most noticeable linear feature type has been interpreted as traces of palaeo-valleys running towards dry gulleys which are running into currently active river valleys, thus showing a potential ancient network of channels suggesting a quite different hydrological scenario in the past. These features turned out to have a marked topographical expression, but it is still unclear which formation processes led to the present geomorphological situation.
Moreover a series of circular and sub-circular features were mapped all across the investigated area; they showed up as soilmarks with a specific spectral signature, suggesting the topographical expression of objects with a brighter runoff top and a darker soil halo of moisture accumulation. After the ground visit, these features have been classified as potential mounds mainly levelled by intensive ploughing to an average height of 0.60 m and an average diameter of 30 m.
Unfortunately the chronology of these features can span from the Early Bronze to the Middle Age and at the moment there is no evidence, apart from their diameters, to assign them to a specific time period.
Finally, the remote sensing analysis helped target the fieldwalking for new potential archaeological sites. In the area investigated, several features which could have been interpreted as archaeological sites were mapped on the satellite imagery, though not all of them with the same certainty. Nevertheless four of the most reliable anomalies returned a substantial concentration of pottery sherds on the ground, highly clustered on the top of each anomaly, thus validating the remote sensing interpretation.
Therefore the fieldwalking season, aside from providing further data in order to improve the understanding of what is going on outside the mega-site within a 5 km radius, made a great contribution to the interpretation of satellite images, thus validating the capacity of spotting new potential archaeological sites while expanding the research in other areas.
This year’s palaeo-environmental sampling consisted of attempts to recover further coring materials from three sites: (a) the alluvial areas adjacent to the mega-site; (b) the site of Lisicha Balka, previously cored in 2009 and 2012; and (c) a salt-marsh site near the modern village of Nova Odessa, some 180 km South of the mega-site.
(a) The Nebelivka N3 core was taken from N48°39’14”, E30°35’41”, at a distance of 3.2km from the Eastern side of the mega-site. The basal level of the core was reached at 4.0 m. The stratigraphy consists of reduced muds throughout with a regolith at base (~3.9-4.0 m)
(b) Although an attempt was made to find a coring site which had a longer sedimentation record than the 2012 Onopriivka core, the coring team was unsuccessful.
(c) Located on the border of the Pontic steppe zone and the forest steppe zone, in which Nebelivka is located, the Nova Odessa – Troitskoe site had been cored for pollen before, being considered as one of key pollen sites for the Ukraine (Artushenko, 1970; Artushenko et al., 1972; Bezusko, 2010). According to existing publications, this site was supposed to be a good indicator of both climate moisture variability and human impact on the vegetation.
When we visited the site and made a detailed survey of the area surrounding the Troitske swamp, we found that in that area there are salt water springs. Local ground water is salinized because of the availability of salt-bearing rocks in the vicinity of the site. The slow dissolution of salt by ground water also was the cause of locally-formed solution sinkholes. One of these sinkholes was most likely the source of more than 10 m of peat as reported by the pollen publications. This is corroborated by published data about radiocarbon chronology; the oldest published radiocarbon date was about 4000 BP (Bezusko, 2010). It should be mentioned that Dr. Bezusko was NOT present on site at the coring campaign and was relying on stratigraphy data supplied by late colleagues who actually did the coring. Our field inspection led to three conclusions: 1) the published conventional radiocarbon age of the Troitske site was correct; 2) the site did form in a solution sinkhole and as such is too young for the purposes of mid-Holocene environmental reconstruction; and 3) the published pollen data of this site are heavily influenced by local halophytic vegetation and not by Holocene climate moisture variability.
This negative result serves as a reminder that potential sites for pollen coring should be looked for closer to the micro-regions surrounding Trypillia mega-sites.
In the event of future pollen research in the Nebelivka area, it will be necessary to discover alternative sites with peat deposits rather than looking for long-distance sites far from the mega-site under investigation.
Since the 2012 excavations of the mega-structure at Nebelivka, a recording team of four (John Chapman, Bisserka Gaydarska, Ed Caswell and Sophia Arbeiter) have paid two visits to the Kyiv Institute of Archaeology for post-excavation studies on the mega-structure pottery assemblage.
During the 2013 summer field season, Sophia managed to complete the recording of the last few boxes of the mega-structure assemblage, before turning, with Ed, to the 2013 excavations by the British side. A similar recording system was adopted, using the same colour templates, photographic recording of decorated sherds and the drawing of all rim sherds.
In all, 835 finds bags were recorded and stored; 8,677 sherds have been fully recorded, 1,274 sherds have been drawn and over 600 photos have been taken
A total of 6,655 sherds, weighing 117.8 kg., was processed from the incompletely excavated Sonda 1. This sample is large enough to be broadly comparable to the mega-structure assemblage studied previously.
A total of 173 sherds, 18 fragments of daub and 7 animal bone fragments was recovered from Sonda 2, where we failed to find the ditch suggested by a linear anomaly on the geophysics plot.
A minimal total of 2,414 sherds were excavated and washed from the 40 test pits. Of these 1,814 (75%) have been fully recorded, producing a total weight of 55.6 kg. This total comprises a series of much smaller assemblages, some containing over 100 sherds while others held fewer than 10; we have yet to make a thorough assessment of the degree to which these groups are representative of the entire assemblages from the houses where they were excavated.
A small fraction of the 2013 assemblage has yet to be recorded. A total of 53 bags of sherds, comprising 816 fragments, has been cleaned and washed but not recorded. This task remains for the coming months.
Intra-site comparisons at Nebelivka (Ed Caswell)
A preliminary comparison of the Sondage 1 (Pit 1) ceramic assemblage (2013 excavations) and the complete mega-structure assemblage led to the following observations:-
The proportion of rims, bodies and bases is largely comparable between assemblages and mega-structure phases. The two assemblages are broadly comparable in terms of decoration, surface appearance, vessel form, colour and presence of burning. In addition, the Pit 1 assemblage maintains the same proportion of rims, body sherds and bases that was seen in each of the mega-structure’s phases, with body sherds typically comprising over 85% of the assemblage, with rims representing c. 10% and the remaining 5% predominately of bases with the presence of only a few complete profiles.
However, the Pit 1 pottery has a far higher proportion of fine ware. This is surprising since, after the first two weeks of excavation of this feature, it was thought that far more coarse sherds were being discovered than had been previously identified from the mega-structure. This suggests that there may be variations in the proportions of fine: coarse wares in different phases of the pit deposition – a variable which will be incorporated into future analyses.
It is also of note that there were far more sherds in the Pit assemblage, despite the smaller area excavated. This will be taken into account when finalising the mega-structure’s interpretation. The Pit 1 sherds found were also typically smaller than those of the mega-structure, with a far lower average, maximum value and standard deviation in weight than most of the mega-structure’s phases. However, this may be the result of a lower recovery rate of smaller sherds from the mega-structure, owing to the variation in excavation techniques over the two seasons.
It is recommended that further comparisons are made between the minimum number of vessels per phase and feature in order to gain a better understanding of the proportion of potparts in each assemblage and the variance in size seen between the mega-structure’s phases and the Pit 1 group.
In general, the 2013 summer season consisted of a series of very successful operations. The coverage of an additional 130ha in the geophysical prospection brought the total area surveyed to 200ha, or 80% of the mega-site. In almost every case, the geophysical plot provided an accurate guide to the locations and types of sub-soil features. The excavations of burnt houses and associated pits produced new and important information for Trypillia settlements, where the investigation of pits has not been greatly favoured in the past.
An important result was that the large pits were excavated on the short sides of the houses, unlike in the case of LBK houses, and proved to be much larger than on the geophysical plan. The filling of the pits relied on ritual action involving the active use of anthropomorphic figurines as well as deposition related to daily life. For the first time, the 2013 excavations defined the initial function of the ‘linear pits’ as supplying building material.
The alternative approach of test-pitting Trypillia structures was successful, with the retrieval of over 250 samples for AMS dating. Soil micro-morphological investigations are in their infancy at Trypillia mega-sites but hold the promise of much entirely new information. Continued intensive, systematic fieldwalking from our heroic team benefited from the detailed satellite coverage of the Nebelivka micro-region but Trypillia scatters continue to remain elusive. The results of the palaeo-environmental investigations were rather limited in 2013 but the cores recovered in 2012 remain to hold out significant promise of a deeper understanding of human impact at the time of the Nebelivka mega-site occupations. Post-excavation recording has been advanced so as to be almost up to date with the 2013 material. In short, much progress has been made towards the Project goals in the five-week summer field season.