Olbia Archaeological Report 2016

Olbia: Trench R-23.

Archaeological report from excavations conducted by the National Museum in Warsaw in the 2016 field season



In 2016, Polish Archaeological Mission “Olbia” of the National Museum in Warsaw (NMW) began its first excavation season at the Olbia site (Ukraine, Mykolaiv Oblast) in cooperation with the Institute of Archaeology of the National Academy of Sciences of Ukraine (IA NASU). The following persons took part in excavation works:

From the Polish side: Dr. Alfred Twardecki (head of the mission); Dr. Piotr Jaworski (trench supervisor); Magdalena Antos (surveyor); Szymon Lenarczyk (photographer, drone operator, surveyor); Maria Bąk (illustrator); Sylwia Groń (illustrator); Dr. Tomasz Herbich (geophysicist); Robert Ryndziewicz (geophysicist).

From the Ukrainian side: Dr. Hab. Alla Buyskikh (head of the expedition); Maria Novychnkova (trench supervisor); senior scientist Viktoria Kotenko (archeologist); senior scientist Yuriy Puholovok (archaeologist); junior scientist Rostislav Segeda (archaeologist); junior scientist Anjelika Kolesnychenko (archaeologist); junior scientist Olga Puklyna (archaeologist); junior scientist Sergey Didenko (archaeologist); laboratory technician I Mykola Sachenko; laboratory technician I Anna Litovchenko; laboratory technician I Volodymyr Selevko; laboratory technician I Olena Malyshevska; laboratory technician I Dmitro Malyshevsky; laboratory technician I Irina Chechulyna; laboratory technician I Andriy Rekeda.

Trainees from the Ivan Franko National University of Lviv (9-20 July 2016): Anastasia Baukova (head); students: Marianna Konoval, Evgeniy Klimin, Danilo Kutnyak, Roksolana Protsaylo, Anna Shvysh, Vera Dutkanitch, Vitaliy Mysak, Roman Kurtch, Dmitriy Kraplya, Rostislav Pelekh, Roman Parkhomyk.

Trainees from the “Odarennost’” School of Fine Arts of the Kharkiv District Council (16-24 July 2016): Mikhail Formin (head), Vladimir Selevko (teacher), Natalya Pyshchulyna, (teacher), Anna Lytovchenko, (teacher), Konstantin Kal’chenko (teacher), students: Aleksander Anan’ev, Mariya Platukhina, Nina Zinich, Bogdan Glushchenko, Tatyana Khmelevska, Bogdan Gayduk, Natalya Zdorovets, Yekaterina Kvasova, Vitaliy Kzymyrovytch, Aleksander Storozhko, Konstantin Itchko, Artem Fedirko.

Trainees from the Berdyansk State Pedagogical University (23 July – 06 August 2016): Valentina Papanova (head), Sergey Bondarenko (teacher), Arseniy Golyk (teacher), students: Yulya Beregovaya, Denis Vykhryst, Aleksey Galitch, Nikolay Ivashyn, Karina Popova, Irina Prokopenko, Elena Sorokina, Yulya Teren, Diana Tchergines, Aleksey Shmatok.

The expedition began on 9 July and lasted until 20 August 2016. Owing to the exceptional nature of the field season (beginning of work at a new site), work conducted this year was preliminary – its main aim was to survey the site, work out procedures related to trench and artefact documentation (all data was recorded in parallel in Polish and Russian) and verify the situation in terms of logistics, accommodation and managing the mission’s operations. All of the above goals have been met. The assumed research objective was to establish 4 grid units (100 sq m in total) and clear the trench from the humus layer and mixed layers of soil, which covered the youngest preserved cultural layers as a result of natural erosion processes. This goal has also been achieved. What should be stressed here is the exceptionally large number of discrete artefacts found (well over 342), in spite of working on a thus designated exploration area. This number surpassed even the boldest initial estimates, allowing us to hope that the selected trench location is indeed situated in a very promising area in terms of further exploration.

Before the start of excavation works, the entire area which was of interest to the Polish mission pursuant to the agreement concluded with our Ukrainian hosts underwent a geophysical survey. Despite the limitations of available research methods, this nevertheless enabled us to establish where to begin field work. Following the geophysical survey (11-17 July), regular excavation work began, which lasted until 5 August. As a result, the base of the youngest cultural layers was reached on the whole surface of Trench 23-R (designation of the Polish trench resulting from the grid of the Olbia site). The last stage, until 18 August, was spent on securing the trench and intensive documentation activities concerning artefacts acquired in the course of excavation work, i.e. several dozen thousand mass objects (mainly ceramics) and 342 discrete artefacts. Mass objects were classified for the purpose of statistical research, and a fraction of them, representing different types of ceramics, was also drawn. All discrete artefacts were described, photographed, and – for the most part – drawn. No drawings were made of coins and certain metal and glass artefacts (e.g. slag); these were only photographed.

What should be emphasized here is the importance of conducting a thorough analysis of all objects recovered from the mixed and humus layers. The last new trench was dug in Olbia approx. 25 years ago – since then, exploration methods have developed significantly, leading to much greater efficiency in acquiring material. Owing to the use of state-of-the-art metal detectors, we were able to recover over 40 coins, including a few silver denarii and ½ of the so-called Olbian ace. It ought to be stated that the first field season already yielded coins that represent virtually the entire operation period of the Olbian mint. We also managed to recover mass objects (ceramics) from the last period of Olbia’s existence (Chernyakhov culture). Following an analysis of this material, we are able to provide a preliminary answer to the long-standing dispute concerning the moment when the Olbian settlement was ultimately abandoned.

It should also be noted that the Polish and Ukrainian heads of missions concurred that more extensive surveying, aided by photographic drone documentation, was necessary in order to create an orthophotographic map of the entire so-called Roman fort (area explored by the Polish mission as well as current and earlier Ukrainian and Russian missions). The lack of such a plan considerably hinders proper interpretation of the architectural remains within the entire citadel and, in a broader perspective, the entire area of the ancient polis. A joint decision was taken to prepare – irrespective of work in Trench R-23 – surveying and photographic documentation of the citadel and, gradually, of the rest of the polis, which would provide a basis for an attempt at reconstructing the borders of ancient Olbia in more detail.


Description and location of the site

The researched area, marked R-23 according to the local grid system, is located within the so-called Roman citadel in the southern part of ancient Olbia, right next to the edge of the promontory cutting into the Boh Estuary, or liman, over several dozen metres (fig. 1; the ancient name of Boh was Hypanis). The lines of Romanian trenches from the World War II are clearly visible along the edge of the promontory embankment. A characteristic landform feature of the researched area are the undulations created by numerous low heaps (measuring a few metres), which are assumed to be the remains of a long-standing practice of excavating building stones by local inhabitants in the modern period.

The location of digging trench R-23 (fig. 2) was carefully selected based on the results of geophysical research and analysing the topography of the site. There are several earlier trenches in its vicinity, out of which trench R-25, located several dozen metres to the south, should be regarded as the most important in light of the current excavation. For many years, it has been the site of archaeological research conducted by a mission of the Institute of Archaeology of the National Academy of Sciences of Ukraine from Kyiv.

Trench R-23 has been marked out so as to form part of the local grid, whose lines follow the north-south and east-west axis. This grid is based on a division into 50 × 50 m squares, which are in turn divided into smaller 5 × 5 m squares. The Polish trench borders with two larger squares, and encompasses four smaller squares, numbered: 210, 211, 230 and 231 (fig. 3). While marking out the trench, we discovered certain inaccuracies of the existing local grid, which were systematically verified and corrected using an electronic total station. Back dirt piles –  separate ones for soil and stones – were to be located around 15 m south of the trench, in the hollow of a former trench.


Archaeological research in trench R-23 in the Roman citadel of Olbia was preceded by a thorough survey of archives and bibliographic inquiries, an investigation of the site and precise geophysical surveying of the research location, so as to ensure that results obtained in the future would meaningfully enrich the current state of knowledge on Olbia’s history.

Before the trench was marked out, the existing local grid was verified in situ using state-of-the-art measuring equipment (fig. 4). A Leica TS06plus electronic total station was permanently present at the trench, and used in topographical and documentation activities.

The total station served to measure the exact location of the majority of discrete artefacts. A DJI Inspire 1 Pro drone with an attached camera proved to be an invaluable asset in topographical and documentation activities, enabling us to take high-quality vertical and angular photographs. Owing to the permanent presence of technically advanced measurement and documentation equipment, it was possible to execute orthophotomaps of the trench at individual stages of excavation work, as well as photogrammetric trench profiles at the end of field work in the 2016 season.

Archaeological excavations were conducted based on the stratigraphic method, and particular care was exercised to preserve the proper order of exploring the existing system of layers. The surveying of elevations in the trench was performed on a daily basis, along with a number of additional measurements. Plans of the entire trench were also drawn up at every stage of the work. All discrete artefacts were additionally measured using the total station and then, equipped with a label containing their description, packed and entered into the inventory. A Fisher Lab F70 metal detector proved to be extremely helpful in searching for metal artefacts. It was also used to regularly analyse the heap, leading to the recovery of around a dozen artefacts. In addition to that, soil from the explored layers was meticulously sifted on an on-going basis. Mass objects, mainly ceramics, were laid out on a special patch, where they were systematically classified, described and counted. Fragments of high diagnostic value were moved to the group of discrete artefacts, and the remaining ones were stored in a separate pile of ceramics.

After the end of the work, the decision was made to secure the trench against devastation by local thieves. Such actions are necessary, as the practice of looting the remains of ancient Olbia is currently widespread, as testified, among others, by the (luckily minor) losses revealed in the central part of the rock rubble on 29 July. The bottom of the trench was secured with foil, and then covered with two layers of loose soil. Nails were scattered on the surface, so as to prevent illegal prospecting using a metal detector.

Descriptive documentation – site notebook, notebook of topographical measurements and notebook of stratigraphic units – was systematically kept throughout the archaeological work. Discrete artefacts, particularly ceramics and glass, but also selected stone and metal finds, were documented using drawings. The remaining finds were photographed. Coins were cleaned and then provided with specialist documentation drafted by a numismatist. All discrete artefacts were entered into a computer database. The study documentation is supplemented with a list of drawings and an inventory book.

Schedule of excavation work

11-17 July 2016 – geophysical surveying

17-18 July 2016 – measurement and documentation activities, marking out the trench, preparing the trench surface for exploration

19-21 July 2016 – excavation work began with the exploration of heaps 1 (SU 1) and 2 (SU 2), related to earthworks most likely conducted in the modern period

22-25 July 2016 – continued exploration of heap 1; humus layer (SU 3) explored on the remaining surface of the trench

26-28 July 2016 – exploring the humus layer completed; exploring the rock rubble (SU 4) in the central part of the trench along the east-west axis;

29 July – 1 August 2016 – continued exploration of the rock rubble

2-4 August 2016 – the level of a late antique layer (SU 5) uncovered on the entire surface of the trench; cleaning the surface of the trench and profiles

5 August 2016 – final photogrammetric documentation of the plan and profiles, measurement activities

6-8 August 2016 – back-filling and securing the trench

A geophysical survey of Olbia using the resistivity and magnetic techniques (2016)

(Tomasz Herbich)

Preliminary remarks and selection of the survey method

The survey was conducted in the Roman camp in the southern part of the polis, in an area limited from the east by an embankment falling towards the bay, and from the west by a back dirt pile created as a result of exploring a trench with an unearthed fragment of barracks. The research was aimed at revealing structures located in the ground which are invisible from the surface level. The survey results would be used to map out excavation work to be conducted in the area by the mission of the National Museum in Warsaw.

Prior excavation work in the locality yielded information as to the scale of accumulations and the type of buildings: the accumulations are at least 4 m deep, while buildings were erected using locally sourced limestone. Consequently, resistivity profiling was selected as the most useful technique of tracing structures, with the assumption that the resistivity of stone walls will be different to that of the surrounding soils. The magnetic technique was used as a supplementary method, as it may serve to record locations that used to generate high temperatures: ovens, hearths and workshops associated with thermal processing. Magnetic susceptibility surveys of trench walls indicated that the material forming the cultural layer surrounding the stone structures was characterized by a susceptibility within the range of 0.30 – 0.50 × 10-3SI. In this context, it is possible that well preserved walls (erected using stone, which is characterized by very low magnetic susceptibility) with copings directly below the surface will be visible on maps showing the changing magnetic field strength – in the form of areas with lower values as compared to their surroundings.

The surface of the surveyed area was slightly undulating (fig. 5). In the northern part is a series of ditches – the remains of embankments created during the World War I (fig. 6).

The survey was conducted between 11 and 17 July 2016 by Tomasz Herbich and Robert Ryndziewicz from the Institute of Archaeology and Ethnology of the Polish Academy of Sciences in Warsaw.


Survey method

Resistivity survey

Resistivity profiling is used to trace the changing values of the apparent resistivity of soils at a constant depth, defined by the type of array used and its geometry. Prospecting depth is increased by increasing the spacing between current probes. The survey at Olbia was conducted using three different arrays, each of which registered changes in layers located at different depths. The twin probe array was used first, with mobile probe spacing of AM=0.5m and fixed probe spacing of BN=3m. An array of such dimensions enabled us to observe resistivity changes at a maximum depth of c. 0.7 m. Sampling density equalled two samples per 1 m2 (one sample per each 0.5 m along lines with 1 m spacing). A Geoscan Research RM15 resistance meter was used in the sampling. The array was moved along a north-south line. For soundings at a greater depth, two asymmetrical Schlumberger arrays were used: an array with probe spacing of AM=5 m and MN=1 m (probe B represented infinity) and an array with probe spacing of AM=9 m and MN=2 m (probe B represented infinity). Mobile probe A was placed to the north of the MN probe pair, and the arrays were moved along a north-south line (figs. 7 and 8). Such an array geometry enabled us to observe apparent resistivity changes at a maximum depth of c. 2-2.5 m (array of AM=5 m) and 4-4.5 m (array of AM=9 m). Sampling density equalled one sample per 1 m2 (one sample per each 1 m along lines with 1 m spacing). An Elmes ADA-05R meter was used in Schlumberger soundings.

Magnetic survey

Magnetic surveying was conducted using a Geoscan Research FM 256 fluxgate gradiometer with a resolution of 0.1nT, sample time of 0.1s, and probe spacing of 0.5m, which measured changes in the vertical component of the magnetic field strength (fig. 9). Sampling density equalled 8 samples per 1 m2 (samples were collected every 0.25 m along lines in 0.5 m spacing) within 20 × 20 m survey squares. In order to enhance the sampling quality, surveying was conducted in the so-called parallel mode (while recording samples, the instrument was only transported in one direction). After sampling was completed within the given rectangular, the probes on the instrument were re-calibrated. Survey lines were marked along the north-south axis. In the event of recording remnants of buildings, if the walls are erected using a material with no magnetic properties, and are surrounded with layers showing a slightly increased magnetic susceptibility – the effective prospecting depth amounts to a maximum of 0.5 m below the surface. If hearths are recorded, the prospecting depth is increased to c. 1 m. In the event of objects that contain a large mass of strongly burnt-out material (e.g., ovens used in the production of ceramics), the prospecting depth is increased to 3-4 m.


Plotting method and presentation of results

Survey results were plotted using the Geoplot 3.0 software. The Despike algorithm was applied to resistivity data in order to remove spurious readings (which were at odds with values recorded in the surroundings). In the case of magnetic surveying, the Zero Mean Traverse algorithm was applied to survey grids where disrupted measurements were caused by difficulties in moving the equipment caused by in situ conditions (uneven mean readings on adjoining survey lines), and the Edge Match was applied to grids where the mean value of readings did not match those of adjoining grids.

The results of the resistivity survey were presented on maps showing the distribution of changes in apparent soil resistivity (further referred to as resistivity maps). On these maps, various soil resistivity values are conveyed by different shades of grey (figs. 13, 14B, 15, 16B, 17ABD, 18CD) or different colour schemes (figs. 10-12, 14A, 16A, 17C, 18AB). In the grey-scale maps, extreme values are conveyed by white (lowest values) and black (highest values). The results are presented in different value ranges: maps illustrating changes in a broad range of values (figs. 13A 15A, 17A) are better adapted to objects characterized by high amplitudes of resistivity changes, while maps illustrating changes in a narrow range (figs. 13B 15B, 17B) are more suitable for visualizing objects characterized by low amplitudes of resistivity changes.

The results of magnetic surveying are presented on a map of changes to the vertical gradient of magnetic field strength (further referred to as magnetic maps), where different values are conveyed by different shades of grey, with black and white corresponding to extreme values (figs. 19, 20). Negative values result from the fact of measuring the gradient (measuring the difference between probes placed at various heights).

Maps on figs. 10-12 have not been interpolated. To render the results more legible, data on maps used to analyse the surveys (figs. 13–20) is interpolated. Readings with a sampling density of 1 and 2 samples per 1 m2 have been adapted to a grid with a side of 0.5 m, while readings with a sampling density of 8 samples per 1 m2 (magnetic survey) – to a grid with a side of 0.25 m. The maps were created using Surfer 8.0 software.



Resistivity survey results

Readings demonstrated considerable diversity of the substrate in terms of soil resistivity on all prospected layers: in the layer measured using the twin probe array (further referred to as the shallow layer), apparent soil resistivity values ranged from 80 to 400 ohm-m; in the layer measured using an asymmetrical Schlumberger array of AM=5m, MN=1 m (further referred to as the intermediate layer) – from 10 to 140 ohm-m; and in the layer measured using an asymmetrical Schlumberger array of AM=9m, MN=2 m (further referred to as the deep layer) – from 5 to 60 ohm-m. The resistivity values in the shallow layer, much higher than in the case of layers located underneath it, result from the dryness (i.e. worse conductivity) of the surface layers.

Shallow layer surveying covered an area of 0.34 ha. The resistivity map shows a number of areas with increased or lowered resistivity values. Increased values may be associated with the presence of concentrations of rubble or gravel and sand; in the case of a site with stone structures, increased resistivity values most likely reflect concentrations of rubble – remnants of destroyed buildings. Anomalies covering the largest area and demonstrating the greatest value amplitudes are visible in the southern (squares D1 and D2, fig. 14) and northern (near the edge between A1 and A2) parts of the surveyed area. A large anomaly with several increased values is also visible in C2 and near the western edge of C1. Some of the anomalies may be explained by the undulating features of the land (which influence such shallow readings). The landform features are at the root of the anomaly in the NW corner of C1 (a hillock is visible here) and the curved anomaly in the northern part of C2, which runs along a fault plane (the area in squares C1 and C2 is c. 1 – 1.5 m higher than the surroundings).

Areas without resistivity readings are visible in the northern and central part of the map. They correspond to the locations of the trench; no samples were taken there, as the landform features would distort all readings.

Intermediate layer surveying covered an area of 0.17 ha. Resistivity maps show oblong anomalies with increased values, oriented along the EW line and measuring up to c. 2 m in width. Such anomalies are visible in squares B1 (fig. 16), at the edge between B1 and C1, and towards the southern edge of C1 and C2. Owing to the shape and value amplitudes of these anomalies, these may correspond to remnants of walls. The map also shows several areas with markedly higher resistivity values as compared to their surroundings. These anomalies are visible towards the western edge of C1, in the southern part of D1, and (an anomaly showing lower value amplitudes) in the western part of C2.

Deep layer surveying covered an area of 0.11 ha. This survey recorded oblong anomalies, oriented along the EW line, whose shape and width correspond to those recorded in the intermediate layer (fig. 17). Some anomalies yielded a more contrasting image than in the intermediate layer – see anomaly located in A1 and near the southern edge of C1 (figs. 17C, 17B). Compared to the intermediate layer image, the scope of anomalies near the western edge of C1 and in the western part of C2 is more limited. On the other hand, structures with increased resistivity values were recorded in the eastern part of C2.

The above-mentioned image of resistivity changes makes it possible to tentatively locate structures which cause such increased resistivity values, interpreted as remnants of walls and concentrations of rubble. The location of such structures is proposed in figs. 18C and 18B.

By comparing two-dimensional images of resistivity changes at individual depths, we are able to obtain information about the third dimension: the depth of structures which demonstrated increased resistivity values.

Since prospecting was done at three different depths, by comparing two-dimensional images of resistivity changes, we are able to obtain data about the third dimension: the depth of structures with recorded increased resistivity. By comparing the image of changes in the shallow and intermediate layers, we see that the high-resistivity structures recorded in the western part of C1 in the shallow layer (fig. 14A) find a clear continuation in the intermediate layer (fig. 11A). The irregular outline of this structure suggests that we are dealing with a considerably thick concentration of rubble (fig. 18C). The oval high-resistivity area recorded in D1 is also clearly visible in the intermediate layer. The maximum of anomalies in that layer is shifted to the south. The shape and value amplitudes indicate that these anomalies correspond to a concentration of rubble preserved at a depth of at least 3 – 2.5 m (fig. 18C; this area was not included in the deep layer survey). Oval anomalies in the western part of C2 are also visible in the intermediate layer (figs. 14 18A). This may also indicate the existence of a concentration of rubble at a considerable depth (fig. 18C).

The comparison also shows that the majority of structures which presumably correspond to walls, visible in the intermediate layer in B1, C1 and C2, are not reflected in the resistivity image of the shallow layer. This means that the structures are present in a layer whose top is below 0.7 m from the surface. One exception are structures recorded in the NE corner of B1 and along the southern edge of C1 – their image in the shallow layer is less contrasting than in the intermediate one. This may indicate that the structures are preserved starting from the bottom part of the shallow layer (below 0.5 m).

A comparison of results from the intermediate and deep layers (fig. 18AB) is much more valuable in terms of reconstructing the depth of structures. Structures recorded in the intermediate layer of B2 are much more distinctive in the deep layer. The same is true for structures running along the edges of B1 and C1 and near the southern edge of C1. The obtained image indicates that the structures are deep and preserved up to a considerable height – at least 2 m. The deep layer image also justifies the assumption that the greater the depth, the smaller the diameter of the concentration of rubble in the SW part of C1, with new concentrations appearing in the southern and central part of C2.

A comparison of resistivity images of layers situated at different depths also enables us to delineate areas with no remnants of buildings (low-resistivity areas). Such areas are visible at all prospecting depths in the southern part of B1, the central part of C1 and the southern part of C2 (figs. 14, 18AB).


Magnetic survey results

Magnetic surveying covered an area of 0.17 ha. Readings demonstrated that the area was characterized by variations in the magnetic fields strength (fig. 19). The magnetic image is hardly suitable for separating anomalies that could be regarded as reflections of buildings. For the most part, changes to the field strength are caused by landform features. An oblong anomaly visible in the NE corner of C1 and the northern part of C2 (fig. 20B) corresponds to a fault plane. The same cause lies at the roof of the anomaly running along the southern edge of D2. The only anomaly that may reflect a wall is the one visible in the central part of C1, whose direction is slightly diagonal with respect to the EW line. This anomaly is reflected in the resistivity image of the shallow layer – it corresponds to a structure which demonstrates higher resistivity values as compared to its surroundings (fig. 20A).

Anomalies visible in the D1 square indicate the presence of a concentration of material with increased magnetic field values. On the resistivity map of the shallow layer, the anomaly location corresponds to a structure with increased resistivity values. This image may indicate the presence of rubble mixed with ashes.

The oval anomaly with high value amplitudes, visible in the SE corner of C1, corresponds to a metal object – most likely a vertical rod (datum point?).




The adopted method of resistivity surveying, which involved prospecting at different depths, made it possible to not only detect anomalies whose characteristic features justified treating them as reflections of remnants of buildings, but also to establish the depth at which they were found. It is also worth noting that the orientation of structures interpreted as walls, and registered in the intermediate and deep layers, is identical to the orientation of walls of barracks unearthed in the trench to the west of the surveyed site (figs. 11–12). This observation clearly strengthens the hypothesis that the aforementioned structured correspond to walls.

The results of magnetic and resistivity surveying of the shallow layer indicate that the only traces of buildings to be expected at a depth of 0.5 – 0.7 m are concentrations of rubble.

The varied soil resistivity and the dimensions of high-resistivity structures recorded provide grounds to assume that the remnants of building may be traced using the ground penetrating radar (GPR) technique. This method is better than profiling in that it yields images of structures in greater resolution and enables to more precisely establish their depth.

Following the end of geophysical surveying, the exploration site was ultimately selected and archaeological work began.

Stratigraphy and dating

In the course of archaeological work conducted in trench R-23 in over three weeks, it was possible to analyse the youngest layers on its surface. Their secondary character is indicated by the considerable chronological span and mixed-up nature of recovered finds.

Heap 1

(SU 1) (fig. 21), found in the western part of the trench, in squares 210 and 230, left over from the practice of excavating building stones in the modern period. This heap was composed of grey, loose soil with a high content of stones (5-15 cm in diameter), and contained high quantities of small fragments of ceramic vessels, broadly dated to 6th c. BC – 4th c. AD. Metal artefacts have also been recovered from the heap, as well as fragments of ceramic building materials (bricks and roof tiles).

Heap 2

(SU 2) (fig. 22), was created at the same time and in the same manner as heap 1, but is smaller and located on the opposite side of the trench, in the eastern part of square 231. The heap is structured analogically to heap 1, although the stone content was markedly lower here.


(SU 3), covering the entire surface of trench R-23. The surface layer is composed of solid, dark, clay soil, which contained vast quantities of highly fragmented ceramics, dated to 6th c. BC – 4th c. AD. Numerous coins were also recovered (5th c. BC – 3rd c. AD).

Rock rubble

(SU 4) (fig. 23), located underneath the humus and heaps 1 and 2, is visible in site because of a small prominence running more or less along the east-west axis and concentrated along the central part of the trench. The rubble contained an enormous amount of stones of varying sizes, combined with clay soil, whose concentration varied in higher parts occupied, e.g., by heaps 1 and 2, and in the central part of the trench. The historical material is mixed up – mostly containing fragments of ceramic vessels and ceramic building materials – and dated to 6th c. BC – 4th c. AD. It may be assumed that the rubble is associated with objects revealed, i.a., below the concentration of stones, which probably resulted from the destruction of earlier stone structures.

Functional layer in the form of a light-yellow, solid clay surface

(SU 5) (fig. 24) found underneath the rock rubble, where we may distinguish the outlines of several objects filled with loose dark grey soil and stones – possibly pits or small basin houses. Large fragments of destroyed ceramic vessels are visible in the solid soil layer, whose exploration, like the analysis of the aforementioned objects, has been postponed to the next season. Historical material recovered in the process of cleaning the top of the layer may be dated to 6th c BC – 4th c. AD, with the late antique material being slightly prevalent.

Description of the finds

Particularly noteworthy is the very large concentration of movable artefacts in the youngest analysed layers. During the first season of the NMW’s archaeological research at Olbia in 2016, in spite of the relatively small volume of explored earth (fig. 25), several dozen thousand mass objects (included in separate lists) have been recovered, 324 out of which have subsequently been singled out owing to their particular scholarly value and entered into the inventory. They represent all periods of the polis’ activity: from 6th c. BC to 4th c. AD.

Ceramic objects represent the largest group of discrete finds (148 in total): these are fragments of serving and storage vessels of various types and forms (83 artefacts), ceramic rollers and amphora stoppers (54), fragments of terracotta lamps (4) and ceramic building materials: bricks and roof tiles (4). The second largest group of finds are metal artefacts (110 in total). The most numerous among them are coins (47), which are described in a separate numismatic report written by Dr. Piotr Jaworski. Out of the (33) bronze objects, the most interesting are: the hilt of a knife (inv. no. pol. 276/2016), fibulae, fragments of arrowheads, a phalera and pincers. Out of the (18) lead artefacts, one that merits mention is a decorative applique in the form of a bucranium (inv. no. pol. 296/2016) (fig. 26). Iron artefacts (12) mostly include nails of various dimensions.

Glass artefacts (35 in total) may be divided into two groups: fragments of vessels (27) and beads (8), while stone artefacts (21 in total) include 9 fragments of marble stelae, including one with a fragmentarily preserved inscription in Greek (inv. no. pol. 55/2016) (fig. 12). Objects made from bone (20 in total) mostly include long livestock bones, serving various purposes and characterized by a smooth surface, as well as astragals and fragments of pins.

The inventory of discrete finds also includes production waste indicating metallurgic and glass-making activity (7 in total) – fragments of slag and metallurgical slag. A piece of red-coloured plaster (inv. no. pol. 261/2016) is also worthy of our attention.

Monetary finds

The first field season of archaeological research conducted by NMW’s mission in the Roman citadel at Olbia (trench R-23) in 2016 yielded a total of 47 coins. With three exceptions, all of the finds are made of bronze and are presumed to have been made by the local mint operating in Olbia between 5th c. BC and 2nd/3rd c. AD. These exceptions are two denarii and one Antoninianus subaeratus (2nd half of the 2nd c. – 1st half of the 3rd c.) from a Roman mint.

The oldest bronze coins from the Olbian mint are cast coins dated to 5th c. BC (11 pcs.) – particularly noteworthy here are ½ of a large ace (inv. no. pol. 36/2016) and 1/3 of a smaller fraction (inv. no. pol. 3/2016). As many as nine finds from that period belong to the group of the famous Olbian dolphins, found both in whole and in fragments. A particularly notable dolphin bears the legend ΘY (inv. no. pol. 81/2016). The majority of Olbian coin finds dated to the 4th c. BC (11 pcs.) belong to a rather widespread group of small coins with an image of Apollo on the obverse side, and a dolphin with the legend OΛBIO on the reverse. Relatively few coins are dated to the Hellenistic period (7 pcs.). A certain increase in the number of finds may be observed in the obtained material only in the case of Olbian coins minted in the 2nd half of the 1st c. AD (6 pcs.) (e.g., inv. no. pol. 72/2016). The youngest local coin discovered in the 2016 field season simultaneously represents the final stage of the history of local minting: this is a bronze Geta’s coin minted at the turn of the 3rd c. AD (inv. no. pol. 267/2016).

It is worth noting that almost a third of the bronze coins minted in Olbia was found in the form of halves, quarters or other fractions. The phenomenon of cutting bronze coins for economic reasons (the demand for small nominal values) is very well known from the Western parts of the Roman Empire, and several, usually very isolated, Greek centres. Olbia will probably also be included in this group, but a separate study is required here. Subsequent field seasons will hopefully bring further finds in the form of cut coins.

Roman coins recovered from the site are quite specific owing to the presence of the Roman army in the area in question. The oldest one is a denarus minted by Marcus Aurelius for Lucilla, which is considerably worn from use (inv. no. pol. 16/2016). Another denarus (inv. no. pol. 201/2016) was minted during the reign of Caracalla, and is virtually free from any traces of use, which could either indicate thesaurization or the fact that it was rather quickly removed from circulation. The Antoninian subaeratus (inv. no. pol. 17/2016) minted by Trajan Decius for Herennia Etruscilla is the youngest among this group of coins – and among 2016 finds in general.

The vast majority of the coins were discovered at the trench location owing to the systematic use of a metal detector. A small part of the finds results from regular sifting of the heap. Since the first research season was devoted to exploring the humus layer of the trench, which contained chronologically mixed up historical material, coins recovered in the course of excavation work cannot serve as a source that would date the archaeological context. Nevertheless the discovered coins represent a coherent contribution to the current state of research on the history of local minting and monetary circulation. They also reflect the specific nature of the citadel analysed by the NMW mission, whose youngest layers are associated with the presence of the Roman army.


The first excavation season of the Archaeological Mission of the National Museum in Warsaw at Olbia was predominantly devoted to a broadly understood investigation of the site. This concerned both aspects related to logistics and documentation as well as the very area where the Polish mission was to conduct its work. Consequently, following geophysical surveying of the site, the decision was made to begin digging in just four squares (100 m2), with a view to exploring this surface to uncover the youngest cultural layers. The latter, in turn, would be explored in the following field season. This way of planning the campaign was aimed at enabling a comprehensive investigation of the specificity of the site, thus minimizing any potential losses incurred as a result of unfamiliarity with the location. Consequently, an unusually high number of movable artefacts recovered in the course of excavation work could serve to smooth out documentation proceedings (field records were kept in parallel in Polish and Russian), so as to meet the scholarly requirements of both the Ukrainian and the Polish side. At the same time, after the end of the expedition, it was possible to estimate the personal resources and equipment required for the following season, where 4 full weeks will be devoted to work at the trench – after cleaning it, the first task will involve the exploration of intact cultural layers.

Although the mixed up character of layers explored in 2016 precludes any attempts at precisely dating the individual stages of using the studied area, we assume that this will be possible at subsequent stages of field work. Nevertheless, several conclusions may already be formulated at the present moment. In the course of the first excavation season, it was possible to analyse and document layers related to the exploitation of stone material in the modern period (heaps 1 and 2; humus). The rock rubble situated below is most likely associated with the destruction of the remains of stone and earth structures related to the last phase of Olbian settlement between 4th and 5th c. AD. The level of land use in the above-mentioned period is probably designated by the surface of the light-yellow solid soil, which shows clear outlines of several objects (pits, perhaps also basin houses). The ceramic material from the last settlement period in Olbia is provisionally dated to the late 4th – early 5th c. AD. As has already been mentioned above, this is the first important research result achieved by the Polish mission. This discovery concludes the long-standing debate between researchers about the end of settlement in Olbia, associating it with the end of the Chernyakhov culture in the area rather than the Hun invasion of 375 AD and the death of Ermanaric, King of the Ostrogoths.


List of illustrations

Fig. 1. General plan of Olbia showing earlier trenches and the location of work conducted by the Polish mission

Fig. 2. General view of the Polish trench

Fig. 3. Location of the Polish trench within the site grid

Fig. 4. Surveyor Maria Antos at work

Fig. 5. Geophysical work in the Polish trench. View from the west

Fig. 6. View of the geophysical research area showing the sampling points

Fig. 7. Resistivity survey

Fig. 8. Resistivity survey

Fig. 9. Magnetic survey

Fig. 10. Resistivity survey results

Fig. 11. Resistivity survey results

Fig. 12. Resistivity survey results

Fig. 13. Resistivity survey results

Fig. 14. Resistivity survey results

Fig. 15. Resistivity survey results

Fig. 16. Resistivity survey results

Fig. 17. Resistivity survey results

Fig. 18. Resistivity survey results

Fig. 19. Magnetic survey results

Fig. 20. Magnetic survey results

Fig. 21. Heap 1

Fig. 22. Heap 2

Fig. 23. Rock rubble

Fig. 24. SU 5 functional level

Fig. 25. Section view of trench R-23

Fig. 26. Bucranium, inv. no. pol. 296/2016

Fig. 27. Marble stela with a fragment of a Greek inscription, inv. no. pol. 55/2016