FIELDWORK REPORT

2003

The present material has been published on cIMeC web-site, in June 2004. The large quantity of information to be processed determined our colleagues to take the short way; hence, although the sent illustration has been published, there isn’t the necessary relation between the text and illustration. Therefore, we seem compelled to make here not a simple link, but to take once again the entire html editing process. We will benefit of this opportunity to make several annotations, from the perspective of August 2005 (exactly, during the third archaeological campaign). The interventions in the text will be marked with the same color as this introductive note.

Archaeological site: Răcari fort (Brădeşti commune, Răcarii de Jos village, Dolj county)

Collective: Eugen S. Teodor (National Museum, coordinator); Dorel Bondoc (Museum of Oltenia); students Magdalena Duţescu, Dan Ştefan (University of Bucharest), Silviu Răduţă (Craiova University)

Period: August 2003
Financing: National Museum

namely 50.000.000 ROL (equivalent of 5000 RON); we make this mention because each report should be judged, among other things, from the point of view of the financing; in our case – a very modest one

The fort from Răcari beneficiated of several tentatives of systematic researches, starting with Gr. Tocilescu (1897-98), continuing with Gr. Florescu (1928-29), who left us a somewhat more consistent report (see Arhivele Olteniei IX, nr. 51-52, 1930). The investigations were continued by the National Military Museum, namely by Cristian Vlădescu, in 1991-92, but afterwards stopped; the repeated tentative to begin once more the systematic research was not concretized until last summer, from the initiative of the National Museum, in cooperation with the Museum of Oltenia. The reasons of re-taking the activity in the area are numerous, mentioning here only Pârvan hypothesis (with “Malvensis colonia”) or the chance numismatic discoveries, including the post-Roman period.

As it may be observed, each of the antecessors has come at Răcari for two campaigns. This coincidence gave us shivers for a while, but from now on, as we are on the third campaign, we can be relieved: the curse has been broken!

As regarding the final episode, with the National Military Museum, I have not understood why this investigation has been abandoned, especially since the three very long sections are, probably in proportion of app 50 %, un-finalized. Possibly – the research would have been re-taken, if colonel Vlădescu hadn’t died, I think in 1995. When we began to be interested of the situation of the Vlădescu “inheritance”, in 1997, we planned the re-beginning of the investigation together with Bobo’s “lieutenant”, Romeo Avram, who meanwhile became lieutenant…-colonel and vice president of the institution; our plans have been compromised by his premature retirement from the military career and, as an implication, from archaeology, in 1999 or 2000. Several years have passed until we have arrived at Răcari...

We will start with several topographic considerations. We agree with the Florescu hypothesis regarding the ancient positioning of the fort in the immediate neighboring of the River Jiu course (up to which, actually, are 1.2 km for the closest point, and over 2 km up to the bridge), from strategically reasons (the fortification must supervise a ford, very probably a military road) as well as from topographical reasons (the terraced aspect of the southern and western sides). Nevertheless, we cannot accept the hypothesis according to which the defence ditch was watered by the river, the difference of elevation mark between the end of the ditch and the Jiu presumptive flowing bench mark being of minimum 7 meters (GPS data). Moreover, the topographic survey (figure 5) performed in May 2003 proved that the defence ditch had no continuity, all around the fort, the access to the gates being made on the soil, not over bridges. Finally, the defence ditch’s bench marks are as well dependent of the placement of the fort on the inclined plateau, the differences of bench marks between the eastern and western sides being of 2 meters.

For the differences of bench marks between the fort and the Jiu River course, the altimetric profile from Figure 11 is instructive.

As regarding the general dimensions of the fort, our measurements, based on a new topographical measurement, are a little shorter than Florescu’s measurements (probably made with the measurement tape), namely 168 (E-V) x 139.8 m (as to 173.2 x 141.5). These dimensions are estimated after Florescu sections, traced along the curtains (!), passing at the axis, at least theoretically, at the exterior of the facing wall. The first section that we proposed intersects the eastern curtain, approximately at the half of the distance between Porta Praetoria and south-eastern corner, overlapping section MMC S.1 from 1991, on a distance of 36 m, from the area of the intersection with S.2/ 1991 until the fence (ruined) of the old archaeological reservation. The decision was motivated by the evident aspect of unfinished work of the military digging, as well by a tentative to estimate the credibility of these researches (partially published, see Revista MMN, supliment 1, 1996). The initial section, 2 m large, was enlarged at 3-3.5 m, due to the action of the medium factors, the new section having the width of 4.4 m, on the axis of the old one.

We will expose the main conclusions of our researches from east to west, namely from exterior towards interior.

The area of the defence ditch presented several surprises.  Firstly, almost certainly, there was a second defence ditch, at exterior, of which our section (1/ 2003) seized only 2.7 m; absolutely sufficient for elucidation (see figure 1, grid units 21-22). The topographic section performed on the slightly inclined plateau from the east (towards Răcari railway station) revealed a level breakage situated further outside the enclosure, which suggests, with enough probability, the existence of a third defence ditch (see figure 2). If this is the case, then the sequences of the defence ditches extends on a depth of 32-33 m, which is understood, given the configuration of the terrain, the eastern side being the most exposed one. The most important of all remains the main ditch, directly neighbored to the curtain, which was the widest and, probably, the deepest. Despite previous reports, which give a width of 8 m (except for the northern side, where it would have been, of unclear reasons, much thinner, namely 4.5 m) we measured 11 m (from the berme to the maximal bench mark towards the second ditch). The depths, as well, are not alike. Measured from the berme, the ditch has a depth of 2.5 m (from the construction bench mark); measured from the exterior, the difference of level between the peak among the ditches and the basis of the main ditch is of 1.8 m. Moreover, despite previous reports, which present the ditch as having an U shape, the section which we undertook revealed a bent V shape (with a sharp middle), the central part, with abrupt edges, having a length of 4.4 m, with differences of bench marks of 1.35 m (towards interior), namely 1.10 m (towards exterior, see figure 1, grid units 18-19).

The filing of the ditch was performed in 5 phases:

a) a slow filling, during the functioning phase, with clayey material, generally without significant organic remains; the base of the ditch probably knew several clearances, more and more superficially, the base of the ditch getting progressively narrower;

b) a filling due to the collapse of the curtain, app 0.4 m of ruins being deposited on the bottom of the ditch; most of the ruins remained on the berme (see infra);

c) massive filling, thick of 1.2 m, with dispersed remains of the curtain wall, most of the filing being represented by earth, with rare materials of archaeological interest; due to the content of the immediate superior level, we must conclude that this phase was relatively short (of several decades), time in which the site was abandoned;

d) a narrow level, of 10-20 cm, with sporadic ceramic remains (probably the beginning of the 4th century), but with a significant organic component, having an aspect of domestic residues, corresponding to the revival of the inhabitance in the former fortification;

At the time of editing this report, we were still under the influence of Florescu report, which presents an impressive list of 4th century coins; today, in August 2005, we believe less in the existence of a 4th century level. This level, of which we talked in the 2003 report, apparently corresponds to the last Roman occupation level, from the discussed segment of curtain, and dates from the middle of the 3rd century. The argumentation will be presented in the 2004 report.

e) an indistinct layer, of “vegetal” type (25-40 cm), due to the last 14 centuries.

The stratigraphic report indicates the fact that the filling process was performed from the interior towards the exterior, the minimal bench marks systematically evolving towards east. An interesting problem raised by our tentative to reconstitute the filling phenomena is the conclusion that the necessary earth volume for filing the ditch was very big, representing approximately 4,85 m3 at each linear meter of ditch, only for the previous marked c phase. Such a mass of earth cannot come from the falling of the vallum. The conclusion breaks down Grigore Florescu’s hypothesis, who supposed that the spoliation of the curtain was the result of the diligence of the modern time peasants, in search for construction materials. For the superior part of the vallum to have arrived again in the ditch, even before the 4th century, is obligatory for the curtain wall to have completely disappeared, until that time. It is very difficult to determine who and why did anyone need, at that historical time (immediately after the abandonment of Dacia), of such a large quantity of construction stone.

The problem had been resolved as well in the following campaign.

As to exhaust the subject of the defence ditches, let’s say that animal bones have appeared under the berme of the stone fortification, which made us to perseverate into digging in an apparent “steril” soil. The perseveration has been rewarded through the appearing of a double ditch, which can belong only to the earth-and-timber phase of the fort. The maximal depth of the arrangement is of 1.35 m, with the observable width of    3.94 m (probably with a total one of 4. 5 m), the first ditch having the width of 2.7 m. The arrangement seems to have functioned for a relatively short time, few organic residues being present at the bottom of the ditch, despite the fact that it must have been very close to the palisade. Two vertical beams have been found, part of the palisade infrastructure, with a thickness of 0.3 m. One was partially over-posed by the foundation, and the other over-posed by…the section of our military colleagues (fortunately unfinalized). The planimetry of the two beams suggests a quite advanced position of the former enclosure, probably with app half of meter. Similar traces have not been found in the interior of the stone curtain. As a consequence, the two defence ditches (possibly 3!) were placed immediately under the palisade, without any interval.

The situation is not unusual; we expect clarifications in section 2.

Note April 2006: earth-and-timber forts without berme exist; we give here only one example from Dacia, such as the fort from Buciumi.

The berme of the stone fort (see figure 3) comprised a row of broken bricks, with a width of 3.4 m, slightly inclined towards the exterior (as not to gain water). At the limit between the berme and the beginning of the defence ditch, the trace of a vertical beam has been observed in the profile, with the diameter of 17 cm and the length of 60 cm, of which 10 cm above the berme. O possible explanation of such an arrangement might be the attempt to stop the falling of the berme in the defence ditch, precaution more necessary as underneath it was the former defence ditch, something that could not be forgotten by the constructors. The compact ruins of the defence wall were found above the berme, with the maximal height of 0.84 m. The ruins comprised career stone of big dimensions (up to 40 cm), but bricks as well, namely big fragments, or even entire bricks (42.5 x 28.5 x 6.5; 40 x 28 x 6 cm). The disposal of the bricks was observed especially at the base of the ruins; therefore it might be supposed that it was due to the superstructure of the curtain, namely the crenellelelations.

Passing to the curtain, we observe that the dislevelment of the actual soil has in the vertical axis exactly the exterior face of the former wall, corresponding thus to Grigore Florescu’s mention regarding the execution of a ditch alongside the wall, on the external side. The exterior facing wall, as well as the interior one, presents the same destruction status (complete, till the foundation, except for several stones from the first row of masonry). Confirming under this aspect the Florescu report, the emplecton, comprising stones of different sizes and qualities, with relatively little plaster, had a somewhat better faith, being conserved on a variable height, of 30-35 m, on a width of 55 cm (smaller on the southern part of the section; probably thinned by Florescu’s research). The spoliated facing walls had the width of 45 cm at the interior, namely 41 cm at exterior. The curtain wall measured, therefore, at the base, 1.41 m, significantly more than in the former reports (1 m at Florescu, maximum 1.13 m at Vlădescu-Avram).

The measurement refers to the foundation, and not to the elevation, of which only the spoliation ditch has remained.

Considering the similar treatment of the two facing walls (see figure 3), our opinion is that they have been spoliated in ancient times, without being the result of any recent activity. What intrigues here is the systematic character of the spoliation, because it is obvious that the respective intervention has aimed not the destruction of the fortification, but the integral recovery of the big stone. Such an action hardly may be considered as one of those who inhabited the site in the 4th century, because, as we shall see, they reused other kinds of materials, at least in the recent investigated perimeter. It is also supposed that the spoliation was realized after the massive ruins from the berme were covered by grass, because from there it would have been recovered, more easily, materials of a similar quality.

Observation pretty unmindful; the ruins are massive, it is true, but comprise few stones of respectable dimensions, and a lot of plaster, which only leaves us to suppose the quantity of stone taken from there.

The foundation of the curtain was made of small stone breaches, river stone and small fragments of brick covered with plaster, the composition being extremely hard. The described mixture is placed on a row of river stones, relatively small and specially chosen (10-15 cm on the maximal dimension). The height of the foundation is of 40 cm, this dimension being as well bigger than those observed by the predecessors.

Immediately at the interior of the curtain, in the southern part of the section (grid units 12-13), we encountered a contradictory complex, at least initially. Parallel with the spoliated facing wall of the curtain and linked with this one, another facing wall was found, comprising two rows of big dimension career stones (up to 40 cm), partially reshaped, of which only the first row may still be observed (see figure 4). The maximal width of this facing wall is of 0.7 m towards south, while the minimal one is of 0.58 m, at north, suggesting a slightly deviated orientation from the direction of the curtain. Towards the interior a spongy stone floor is arranged (local yellowish sandstone, named “sâgă”), flat, floor crossed at app 2 m towards the interior of the fort by a late arrangement. All the elements indicate a tower, crossed at north by the section from 1991, at west by the subsequent arrangement, and at south limited by the very edge of our section. Thus, for the moment no complete dimension is known. The tower stratigraphicaly superposes the first clay levelment (foundation levelment), but it is crossed, above the first row of masonry, by the construction levelments of the curtain. The objective cannot be therefore ascribed to the stone phase of the fort, but to anterior arrangements. It must be observed that the tower is placed exactly at half of the distance between the SE corner and the southern corner of Praetoria gate (see figure 5), which corresponds to the general rules for construction of a 2nd century fort (see Drobeta). The discovery of a tower foundation in exactly the expected position suggests, as well, that the plan of the earth-and-timber fort was, at least in general lines, repeated by the plan of the stone fort, although, in this particular case, the intermediary tower was not built. Obviously, the stone foundation does not necessary implies a heavy structure, being more plausibly for the structure to be of timber; the stone foundation was necessary in order to sustain in good conditions a timber storied construction, of minimum 6 m height.

Situation symmetrical for both profiles, the foundation level was crossed by an arrangement which succeeds in time the construction of the curtain wall (see figure 3), at approximately two meters towards interior, over which the earth taken from the defence wall was deposited, clearly distinguished from the stratigrafical point of view of the earth from the surface, blackish, and of the one from the depth of the ditch, with native clay, practically without anthropic traces (isolated). The separation angle between the two categories of filling suggests the fact that the earth was thrown from a high altitude, therefore after the curtain wall was build, at least only partially. The deduction is confirmed by the stratigraphic situation from both sides, under the earth from the agger being present several strips with plaster traces and finely broken stone. On the southern profile, it may be seen another clear suggestion of the moment when the vallum was raised (and the defence ditch dug): at the base of the yellowish earth is present a group of broken bricks, suggesting that, at the same time, the crenelleles were built.

This was one of the most difficult problems raised by the first campaign, and two years were necessary in order to offer a satisfactory explanation. We have been clarified by the lecture of several passages from professor Isac’s book regarding the earth-and-timber fort from Gilău (Dan Isac – Castrul de cohortă şi ală de la Gilău), where it is emphasized, in several situations (p. 28, 33, 41) that some wood constructions were “daubed with white plaster”, with thickness of up to 1 cm. If we are taking of plastering (applied on wood!) we are therefore in the situation of a material which we today call “polishing plaster coat”, namely a lighter plaster.

Note fall 2008: in the next campaings the light plaster was repetedly associated to the eart-and-timber fort, interpreted as a protective coat for the wood. As regarding the two-colours agger, the first intuition was faulse. Both layers of earth in the agger belong to the stone fort (which has two phases). No part of the earlier rampart of the second phase (eart-and-timber) is preserved, as concluded after the campaing from 2006 (see report).

The purpose of the arrangement that crosses the foundation level could not be explained; this arrangement, certainly in relation with the construction of the fortification (both of the profiles, placed at a distance of 4.4 m, are identical); furthermore, something similar appears on the single fragment of profile published at the Military Museum in 1996, and which refers to the same section 1, but to the western curtain, opposite to the discussed position. Provisory, we may consider that it is a drainage ditch, as to limit the watering phenomena (favored by the large claying process made with the occasion of different arrangements).

As regarding the initial shape of the vallum, we can not adhere to Florescu’s opinion, according to which this would be practically identical with what may be seen today, because he would have observed an “arrangement with little stones” (?). In the superficial area, traces of different construction materials may be seen, but not because there would be a Roman settlement, but due to the spoliation and/ or levelment activities, the later of recent date, namely the beginning of the 20th century, “for giving back the earth to agriculture”. The reason is very simple, and has already been exposed: almost 5m3 of earth for each linear meter, has fallen from the agger back in the defence ditch. It is pretty simple to say that this considerable mass of earth fallen from somewhere, namely from the agger. Even if we would prolonge the observable line of the vallum, up to the intersection with the curtain, as in figure 6, the mass of supplementary earth is of only 2.4 sq.m. at the profile, namely 2.4 cubic.m. at linear meter. From this working hypothesis, for the curtain would result a height of 3 meters, plus approximately 1.6 m the height of the crenelle, which, for that time, was reasonable. There would be no other round road except for the curtain itself; we do not exclude the possibility of a supplementary arrangement, of timber, which might have enlarged the round road. Finally, we can not remark that other 2.4 cubic m. at linear meter are missing from the calculations, which might suppose an even higher vallum, but since we have no analogies, we stop here.

Nicolae Gudea (the Roman fort from Buciumi, fig. 9, p. 92) proposes a very similar reconstitution, with the aggradation of the earth vallum up to the step level on the wall (therefore, at app +3 m from the foundation). The difference is that he has aggradated the vallum under an angle which to allow the organization of a round road as well, sideway of the wall. I have tried to respect the angles suggested by the stratigraphy, and the result is the drawing from figures 6 and 14. Nevertheless, we can not exclude a timber structure, at 1-2 m behind the wall, capable of hedging a larger amount of earth, which would allow the organization of a round road slightly wider, on which two armed persons to pass without any inconvenience.

As it has been already observed at figures 3 and 6, at the limit of black earth filling with clayey filling, of the vallum, there is an isolated complex. It represents the rear back of a 4th century kiln, of usual type, which appears together with the settlements of this century. Of the settlement nothing could be recovered, the military section from 1991 being performed exactly on the middle of it; the ruining of the edges done the rest, therefore we had to compel with several ceramic fragments recovered from the centre of the section, in grid unit 10, area in which the complexes over which we shall discus later, were superficially deranged. A lid pot detaches of these ceramic fragments, piece very rare in the epoch.

Next to grid units 9-10 the military archaeologists stopped the investigation at a depth of 20 cm, on compact debris, which they compelled to name “via sagularis”, for the sole reason that, planimetrically, it should be there. The area from grid unit 10 towards west did not represent a priority for 2003 campaign, being obviously that there won’t be the necessary financing for an exhaustive investigation of the area, offering thus priority to the fortification area. The ruins clearance proved to be a complicated task, with a lent research evolution, which determined us to work only on the northern half of the section’s width (see figure 7). The ruins proved to belong to a 4th century settlement arrangement. After the ruins’ removal, we recovered an enterable incised tegula mammata (quadrate, with a side of 43 cm), broken in situ (figure 8). It represents the tegular material recovered and reused for other means than the original ones, very probable as interior furniture.
In grid unit 9, immediately under the vegetal, appeared a double brick facing wall, oriented approximately parallel to the section. The wall from the edge of the section is made of two or three rows of bricks, without binder. The bricks have different sizes, including quadrate, without doubt recovered from a heating system. The wall from the interior of the section has the bricks placed on the edge. According to the appearances, the two brick facing walls represented only the sides of an earth wall, after the model of a wall with a military function. The thickness of this arrangement measures almost half a meter (0.47 m). The settlement certainly is posterior to the abandonment of the fortification, not only because the ceramic material belongs without doubt to the 4th century, but because the arrangement superposes what our colleagues have rightly suppose that should be: via sagularis (figure 9). We also mention that the building system of the late settlement would seem to be completely “barbarous”, if it wasn’t analogue to the construction system of the military barracks from the interior of the fort and from its functioning time. Near the grid unit 8, the wall seems to make an angle towards north, which might suppose that those bricks placed on the edge represent the exterior of the building, and the remainings of a similar arrangement, from the centre of the section, would represent a second settlement complex; from the same grid unit 8, at only 30 cm towards south, another brick placed on the edge was recovered, just 30 cm lower, oriented along a clay facing wall, belonging to the 3rd century.

We have to retain and to avoid the expression “without doubt”; the discussed settlement represents the final Roman occupation phase of the area, but dates from the middle of the 3rd century; the “lower” brick from the commented passage belongs to the final phase of military barracks.

Via sagularis comprised river stones of small dimensions, similar to those underneath the curtain, and very rare fragments of other construction materials, on a single row, without infrastructure. We counted at least four re-buildings, without completing the investigation, made by simply adding a row of river stones over the mud dipped road. The width of the road is variable; the oldest variant that we could measure has only 2 meters, but the re-buildings seem progressively larger, displacing the clay from the ditch and thus arriving at 2.65 m. We would have expected that the road to be edged by a draining ditch, at least from the agger, but such an arrangement could be observed neither in the ground, nor in the profile. As regarding the convex shape, typical for the Roman roads-no trace, the levels being even slightly centrally lowered, due to the inexistence if the infrastructure. From the little archaeological materials recovered from via sagularis, detaches a bronze coin from Gordian and a pilum, both of them recovered from the latest level of arrangement.

Wrong; the road also had a ditch, towards east (curtain), but it would be observed only in the following year.  

Westwards the road, linked to this one, was discovered a construction with military purpose, but the functionality of which cannot be, for now, determined. Some of the small objects discovered (bronze wire, slag in several points and on both of the levels) would suggest a reparation workshop for different metallic objects. The walls are built in the described technique, with two brick facing walls, with or without clayey bound, with clay filling (figure 10). Without the system to be observed as clearly as in the 4th century arrangement, it was easily to observe the appreciable frequency of the bricks along the clay facing wall, very well preserved, although on a height of only several centimeters. The most interesting case was the one of a 3 superposed bricks package, from grid unit 5, near the southern side, the position of which (arranged as playing cards and rotated to 90° towards the wall) suggest that they have fallen from bigger height. The direction of the bricks dispersion suggests as well that they could have fallen from the eastern wall, placed at a distance of 2 meters, and not from the western one, situated at only 40 cm. If the observation and deduction are correct, then the group of three bricks fall from an altitude of at least 2 meters, which allows us to suppose that the wall was built, up to the top, in the same manner. The three observed walls are not actually parallel. The one that borders via sagularis is inclined, as to the section, with almost 7°, while the deviation of the others is much smaller, under 1°. Grosso modo, all of them are parallel to the curtain, which presents a deviation of 3°. Despite these lineament differences, the common sense, as well as the stratigraphic situation and the inventories, recommends the interpretation that all the three spaces delimited by walls are the rooms of the same building. The distance between the walls is, from east to west, along the southern side, of 3.23 m, 3.05 m, namely another space of 1.64 m up to the border of the investigated area (the first three grid units, projected initially for the intersection with S.2/ 1991, were not performed due to the lack of time). From via sagularis until the western border of the section are 8.57 m. The width of the interior construction should be, according to the canons, of 10 m, which is perfectly plausible.

Slag has also appeared in previous campaigns, highlighting the professional function (of “reparation workshop”) of the arrangement from the superior level.

Of the inventory elements recovered from the ruins of the barrack, we mention a key, knuckles and construction bolts, a cachet handle, a fragmentary iron buckle, a bronze card, elements already considered in connection with the presumptive professional activities of metal processing, glass fragments, a terra sigillata fragment from an earlier level, but as well more surprising elements, such as a fragment of bronze bracelet, a bead and even a fusaiol (on a later level). As may anyone ascertain from the presented documentation, the investigation is not finalized in the area of the interior construction (barrack). Even though, we may assert from now the existence of two absolutely distinct construction levels, on the same position, the ending of the older phase, as well as of the newer one, being a brutal one, with burning traces. What must be added, as concerning both of the constructive phases, is that the number and the fragmenting of the discovered tiles do not sustain the idea of a covering with tegular material, but with lighter materials, that would have favored the burning.

The idea will be corrected, in the following campaign, after we have counted the number of fragments; at least the new level of the barrack was, with certainty, covered with tile.

The 2003 campaign offered the possibility of topographic activity evidence around the fort, with the aid of GPS technology. The tracks of the civil settlement may be traced 200 m towards W, SW and SE of the fort, the archaeological situation being undoubtfully determined by the configuration of the terrain, which goes down, on these directions, at a level which actually means easily flooded terrains. Towards E the traces of inhabitance seem to exceed 300 m. Towards north, most of the settlement is compromised by the construction of the railway, of the road and of the modern settlement. South of the railway, the Roman settlement exceeds, anyhow, on approximately 800 m on the axis NW-SE, and approximately 400 on the axis NE-SW, which means a perimeter of approximately 30 ha (including the fort). The topographical study has confirmed older and newer opinions regarding the possibility that, in antiquity, the course of Jiu River to have been much closer to the fort, namely 200 m SE, as the closest point, and approximately 300-400 towards SW (see figure 11). One of the most interesting aspects observed during the surface investigation is represented by a vallum, still visible on several meters, as the north-western limit of the settlement.

According to the next year financing, we propose several immediate objectives, with the following priority order: the finalization of the first 10 grid units westwards, in order to obtain a complete stratigraphical sequence; the prolongation of the section towards east, as to intercept the defence workings placed at the exterior of the main ditch and of the reservation, as well to intercept their relation with the civil settlement; the development of the investigation towards interior; the specification of the dimension of the civil settlement through surface survey and GPS mapping of the types of archaeological material present at the surface.

See as well figures 13 and 14, attached to the report at the same time with the preparing of the on-line version.