Issue 1, 1999
Histria's New Look - a 3D Model

by Mircea Angelescu

Mircea Angelescu, Histria's New Look - a 3D Model , Living Past, 1, 1999,

Table of content

Keywords: Histria, method, 3D modelling, aerial photos

An essay of 3D modelling restitution of the cliff on which the Greco-Roman site of Histria is lying. The method used is presented and the results for the main areas of the city as well.

Collecting and recording archaeological data is still made by the traditional methods, whose evolution in the last 50 years is scarcely observable. The nowadays technological evolution, mainly in the computer's field, made possible that in many countries the archaeologists developed new recording methods, in order to offer important supplemental information for the research. Recording plans on millimetrical drawing paper and the materials lists in the excavation notebooks is very useful but it is slow and does not permit but a limited treatment of the data resulted by excavation. The actual method is allowing only an indirect connection with the Click to see the plate topographical or/and geographical data, too. For instance, in the best case, a pottery fragment is placed in its archaeological context but there is no possibility of connection with its spatial co-ordinates, referring to the excavation area or to the whole site. With no exception all the archaeological data recorded during the excavation are using -- when they are registered -- the two co-ordinates planimetrical system.

This system is not only limited but - practically - not very precise mainly due to the fact that the reference system is based on non-fixed points that are susceptible to move during a couple of years, too. The next step, of assembling a general plan is, consequently, imprecise and is producing non-acceptable errors. If the archaeological site is a large one, if on this site there are several simultaneous researches which are separated by large distances, assembling all the data in a single plan become practically impossible not only because the enormous data quantity that should be correlated, but because of the cumulative errors generated by the different excavation points, too. One of the main limits of the actual method, evidently resulting, is that it does not permit spatial distribution analysis, nor the statistical analysis of the data resulted of the archaeological research. It is obviously necessary to adopt a new recording and managing system of the archaeological, architectural and topographical data, which will allow to correct the actual errors of this method. The three dimensional modelling and the GIS systems (often based on the GPS) are today used on al large scale by archaeologist all over the world. The first method aims to offer a "solid" entity that can be manipulated (rotated, scaled,zoomed) in order to provide a better "vue" of the remains or of the soil. It also offer a solution to one of the paradoxes of the archaeology: the archaeologist can not see in the same time all the results of his excavation, though archaeology is often defined as a mainly observational discipline. The second method is used not only to visualise different sets of data (level maps, plans, aerial photography, excavation plans etc.) but also allow statistical and area distribution analysis due to the graphics-database approach.

What follows is the description of the first stage of a scientific research project which is proposing the modernisation of the recording and managing archaeological data method on the site of Histria, the world wide known Romanian archaeological site, not only as a national archaeological-school but as a pioneer of the new methods, too. We applied the method that is known as three dimensional modelling of the surfaces, a method that is offering, among other advantages, the possibility of visualisation of surfaces that the planar method is not able to offer.
Click to see the plate  The first stage consisted in topographical (i.e., level) measurements, in those points in which the cliff can actually be seen and where the archaeological excavations have reached its level. All the level measurements were made by reporting them to a "0 point" that is the geodesic mark placed on the site -- the results being in this case negative figures. The height of this point, compared to the sea level, is 8.70 m. All the data was then put in a database for an easier manoeuvrability [1]. The second stage was to place the measurement points on the plan of the Late Roman city and to build a rectangular grid covering the whole surface. The result was a total of 10,000 rectangles (100x100) their corners being the surface determining points. The points with unknown co-ordinates where determined using the Krig algorithm (more precise then that of the polynomial regression or that of the inverse distance of a power) for the interpolation, statistically probable, of the other point's co-ordinates [2]. This was the method used to determine a 3D surface for each of the 10,000 rectangles, perfectly placed in space -- and finally the total surface assembled from the 10,000 smaller surfaces. We have transformed the result of this research -- that is the modelled surface of the cliff -- into a three-dimensional entity that was "treated" in order to make it look as possibly close to the reality and to permit us to "look" at it from any desired point. Click to see the plateTrying to obtain a virtual image of the whole site, the cliff surface was superposed by a drawing level containing three-dimensional structures (mainly walls and monuments). The monuments were drawn at the1:1 scale --using the centimetre as drawing unity. The solution found to the major difficulty of the spatial positioning was to scan the city's plan and then to overlay the result over the initial drawing -- this method offering us the validation of our results. An important mention: our aim was reproducing the volumes of those structures and not the details (we were mainly interested in respecting both the dimensions and the planar and spatial co-ordinates of the walls). Each of these 3D entities was automatically recorded in a database used to verify or edit whatever information needed: the distance between two arbitrary points/ length, width, height, perimeter, area, volume calculations, any planar or spatial co-ordinate of any point of the model. The next step was tot digitise the plan of the city and the topographical map of the site. The level map was realised in 1953 by officers of the Military Topographical Direction under the scientific co-ordination of the architect Dinu Theodorescu and it used level curves for each 0.5 m. The main difficulty was that this map contained only the planar co-ordinates of the Late Roman city wall and level figures that were related to an unknown reference point(s). Thus, the lowest level (i.e., on the coast line) was labelled 15.5 m. and the point for the geodesic mark was labelled 22.95 m. We had to verify the relative differences between those figures and it appeared that the figures were correctly registered. After converting the heights of the level map with the sea level height of our "0" point (22.95 m. - 8.70 m. = 14.25 m.) we obtained a level map using absolute figures for the heights. Click to see the plateThe result was a three dimensional representation of the actual soil level. This plan was also divided in 50 Click to see the plate x50 m. squares that will allow future excavations to be accurately located on the plan that is covering now the whole site (750 x 1200 m, i.e. 24 x 16 squares). Each of this surfaces is divided in 10 x 10 m. squares and they are offering a high precision grid for any excavation that is to be registered in the future. As base- point it was chosen the south-western corner of the Late Roman city gate that by its solidity is guaranteed to be a fixed point for a long time from now on. Choosing this point also permitted us to combine and verify the superposition of the plan with the level-map at our disposal offering a fixed point for the orientation of the complete plan.

[1] We used Autocad r.12, trademark of Autodesk Co.

[2] The software used was Surfer 5.01, Golden Software Co.

Aerial photographs

An other verification resulted by comparing our results with the aerial photograph made in 1981. The only problem that we had was that the coast line had not the same tract as that of the aerial photograph. In this context it is also interesting to underline the fact that comparing the first aerial photograph of Histria, published by M. Lambrino in 1938 [3], with the level plan of 1953 and with the aerial photograph of 1981 it appeared that the coast line of the Sinoe lake has modified - it retired with 50-60 m., as shown in the table below:

[3] The aerial photograph must have been made before 1935, as shown by Al.S. Stefan, Revista Muzeelor si Monumentelor, 1975, 2, p.57, n.66.

DateDist. 1Dist. 2Dist. 3(All distances are approximated)
193485 m.65 m.75 m.Dist. 1 = distance Tower A - Sinoe lake shore line.
195360 m.42 m.50 m.Dist. 2 = distance "fracture" late Roman city wall (East) - Sinoe lake shore line.
198127 m.13 m.17 m.Dist. 3 = distance Pârvan's sondage for the east side of the late Roman city wall (that correspond to a point between the C and D towers on the western flank of the late Roman city wall) - Sinoe lake shore line.

These results have three major implications:
Click to see the platea. - interpreting the descriptions of the site given by the reports of Pârvan and Lambrino must take this fact into account.
b. - the difficulties encountered in locating the excavation conducted by Pârvan [4] in the area called today "Temple Sector" are also due to the modification - in this area - of the coast line with more than 50 m., i.e. most of his 200 m2 area must be now covered by the lake.
c. - the reported "wall" or "dike" that would be some 50 m to the east of the coast line - now covered by the water of Sinoe lake - roughly correspond to the old coast line. Even the presumed orientation of this structure (approx. NE-SW) could be explained by the old coast line that look like there has a small gulf just in face of the Temple sector (the second gulf was placed on the other side of the "Domus" promontory). For the final result we preferred to use the level map - with a slightly corrected coast line - because it offers an easier way to control planar co-ordinates. All these difficulties and results gave us more reasons to intensify our efforts for implementing at Histria a GIS intended to cover the whole area of the ancient site.

[4] V.Pârvan, Raport provizoriu asupra primei campanii de sãpãturi la Histria, in Anuarul Comisiunii Monumentelor Istorice pe 1914, Bucureti, 1915, p. 19,22 and 26.

Main results
We shall now, very briefly, consider the main results of this research for the space usage in the main areas of the site.


Click to see the platePerhaps the most surprising and important result of this research is concerning this area, placed in the north-eastern part of the site. Pârvan's presumption about the existence of Greek temples led to an excavation in the area. After WW2 the archaeological researches conducted here confirmed the presence of one of the sacred areas of the site. Unfortunately, no information survived on the location and results of the old excavations. There is no way that we could identify the placement of this excavations as we are unable to locate on the terrain the city wall "fracture" mentioned by Pârvan in the area [5]. However, the re-discovery of the Late Roman city wall in the course of recent archaeological researches and its supposed traject seem to confirm that the "fracture" is to be placed in the area immediately to the east of the altar D. Future researches in the area are supposed to uncover the late Roman city wall on a longer distance to the north [6]. The main result of this research is that the three dimensional model of the cliff put in evidence,too, the fact that the "Sacred Area" is not on the Acropolis. We used to consider, because of the general spread idea of the usual placing of the cult monuments on the Acropolis of the Greek cities [7], that Histria's sacred area is not an exception to this rule and most of the publications are referring to it as being placed on the Acropolis of the city [8]. As our three dimensional model is showing, the platform surrounding Aphrodite's temple is approx. 3 m lower than the highest area of the city placed some 50 m. to the south. The questions that first come to one's spirit is: what have determined the placement of the "Sacred Area" here, in one of the lowest areas of the Greek city ? In our opinion, an acceptable explanation of this situation could be the deep depression located immediately to the east of the platform of the Aphrodite's temple. Of course, the sacred character of this cavity is only to be supposed and the lack of stratigraphical information contained by the publications that refer to the excavations made in this area are not making things easier. As our 3D model is showing, the shape of the cliff could be an answer to the questions risen by the non-continuity of the use of the sacred space at Histria: there is no trace of superposition of the Greek sacred monuments by early Roman temples and perhaps the Roman temples - and other Greek temples as well [9] - are to be found on the highest area of the city.

[5] See C. Domãneantu, A. Sion, Incinta romana tirzie de la Histria.Incercare de cronologie, in SCIVA,1982 (4), pp. 377-394, for a comprehensive account of the researches on the late Roman city wall and the bibliography of the subject.

[6] We wish to thanks to Al. Avram for the permission to collect topographical data resulted from the recent excavations.

[7] Of course, temples are not to be found only on the Acropolis of Greek cities. For few of the many examples, see W. Hoepfner and E-L. Schwandner, Haus und Stadt im Klassischen Griechenland, p.145 (Priene) and p. 188 (Halikarnassos).

[8] E.g., Histria I, p. 231, p. 20, P. Alexandrescu, Histria in archaischer Zeit, Xenia 25 (1990), p.52.

[9] M. Lambrino, Les vases archaiques d'Histria, 1938, pp. 356-357, supposed the existence of a temple dedicated to Apollo

Click to see the platePlaced on the highest area of the site, this late Roman habitation quarter was - most probably - inhabited from the beginning of the Greek settlement. There are three late Roman houses (Domus I - III) placed around one of the most important street intersection of the city, but the archaeological excavations conducted here are still unpublished and the most of the data resulted is lost. The excavation - in course - of the fourth house (Domus IV) is intended to recover as much as possible of our knowledge about the area. Aside of the main excavation small sections were made, mainly across the streets, attempting to relate the stratigraphy of Domus I - III with that recorded for Domus IV. The sections S. I, S. II and S. III made by Octavian Bounegru in this area had as one of the results the observation that the intense decapations have been made in the Acropolis area during the Hellenistic and Roman ages [10]. The results of the excavation conducted by the author in the excavation area of Domus IV - during the 1995 campaign (S. IV) - put in evidence a similar situation: a layer containing pottery fragments of the first half of the VIth c. BC was simply superposed by a layer containing the ruins of the VIth c. AD house. The situation need to be explained. This section was traced perpendicular on the supposed traject of the late Roman city wall and was expected to offer stratigraphical data that would allow the connection of the archaeological complexes of the city wall with those of Domus IV. It appeared that the city wall was following the natural limit of the rock that is higher with approx. 7-8 metres than the lower area lying to the south of the city. The natural form of the terrain and the modern interventions altered the stratigraphical data. The main result for this study [11] is that in the construction moment of the house of VIth c. AD a decapation destroyed all the remaining of the precedent constructions. The depth of this intervention is given by the undisturbed VIth c. BC layer that remained intact. It is clear that the shape of the cliff (higher here than in the other areas) had imposed this solution for the constructions erected close to the natural southern border of the acropolis. The Greek and early Roman architectural monuments supposed to be on the Acropolis should be searched, and probably found, rather in the central area of the city.

[10] The results will be published in the near future by M. O. Bounegru.

[11] The very interesting results that this excavation produced will be published in the full-detail report.


Click to see the plate As our measurements are showing the central area of the late Roman city is higher than the rest of the site. Of course, the continuous inhabitancy of the area since the beginning of the Greek colony led to an important accumulation of anthropologic deposits - approx. 6 m. high. On the other hand the excavations conducted in this area didn't descend, but accidentally [12], under the layers of the VIth c. AD. The main monument uncovered in this area is the impressive Bishopric basilica - built during the reign of Justinian. **trimitere la articol Dacia Suceveanu** The dimensions (length = 67 m., width = 30 m.) and the monumentality of the remains of this basilica will not allow further researches of the lower layers - with the exception of some small size sondages. One of the sondages was made by the author of this study. The result was that it permitted the observation that the exterior walls of the basilica descended and rested directly on the cliff. From the stratigraphical point of view the situation can be summarised by the fact that there were identified all the Greek and Roman layers known from the excavations made in other points of the site, from the Greek archaic age down to the last Roman layer of the city. An other important result, despite the reduced size of the sondage (4 x 2 m.) was that no trace of decapation could be observed in this area the layers succeeding in the most natural order. Of course, further sondages that will be conducted in connection with the excavation in course of the Bishopric basilica will bring new data that will help having a better image of the buildings lying under this monument. Our sondage did not offer any indications of important Greek buildings, but its size was - evidently - not permitting the observations that could be made in the case of a surface research. If we accept the idea that the Acropolis (in its urbanistical acceptation) is to be identified with the highest area of the cliff , found in the centre of the late Roman city, then we should suppose that the area surrounding the Bishopric basilica (which is superposing an earlier basilica ) could -- in that case -- overlay some of the main monuments of the Greek city. According to the "law" of the continuity in the urban space usage, if the Bishopric basilica of the VIth c. AD is built on the same place as an earlier (and smaller) basilica of the IVth c. AD we should suppose that these monuments were built over other older religious buildings. It is evident that this hypothesis is to be confirmed by future excavations.

[12] O. Bounegru, in La politique édilitaire dans les provinces de l'Empire Romain, Actes du Ier Colloque Romano-Suisse, Deva, 1991, Cluj, 1993, p. 195 - 200.


Click to see the plateIt is now obvious that the "Pârvan" basilica is placed on the southern border of the Acropolis. This could be an other reason for which Ms. Lambrino's theory on placing here Histria's most important and greatest Greek sanctuary - the one dedicated to Apollo Ietros - is to be reconsidered. If - as it is now presumed - the port was placed somewhere to the south of the city its position would have been as close to the port as close was Apollo's temple in Histria's mother city - Miletus [13]. The same placement of the Apollo temple, on the border of the sea is encountered in Rhodes [14] Also, it would be a very interesting situation because we would have - for the first time at Histria - a Greek temple superposed by a Roman basilica. Of course, an ultimate confirmation is to be found only by re-opening the old excavations and, thus, verifying the conclusions that led M. Lambrino to her conclusions.

[13] For the placement of the milesian sanctuary dedicated to Apollo Delphinion on the border of the Lions Harbour see W. Hoepfner and E-L. Schwandner, op.cit., p.9 and 11.

[14] W. Hoepfner and E-L. Schwandner, op.cit., p.23,


Click to see the plateThis area is much lower than the rest of the late Roman site. Even lower seems to be that quarter of the city that was attached during the IVth c. AD (phase II B of the Roman city). The ramp of the streets of this area seem to follow the shape of the cliff . The excavations conducted here stopped on the late layers of the occupation and show no trace of any structure that would have marked somehow an urbanisticalal distinction between the old quarters and the new one. However, the shape of the soil is quite *** here. We should add that it was demonstrated - by recent excavations - that this "economical" quarter is lying directly on the ancient sand of the gulf existing at that time to the south of the city.


Click to see the plateThe western front of the late Roman city-wall is placed outside the high area of the city. This part of the city wall is placed over the Hellenistic layers [15] - which demonstrates that in the Hellenistic age these layers were lower than the rest of the city. Prof. D. Adamesteanu informed about the existence - in the civilian basilica placed just across the city-wall tower G - of a huge wall, running parallel to the tabernae walls, found at a depth that was making him think it was Greek [16]. If there are urbanistical reasons [17] to presume the existence in this area of a wall the same is to be said about the topographical reasons: if the wall exists it is most probable to follow the ridge of the high area of the centre of the Greek city. The supposed urbanistical function of this wall - assumed by Al. Suceveanu [18] - is still to be revelled by the planned excavation.

[15] Gr. Florescu, Histria I, p.119.

[16] Partially the same information at Gr. Florescu, Histria I, p.115.

[17] Trimitere Monica

[18] Al. Suceveanu, Histria - bilant si perspective??, in SCIVA, p.??????


Click to see the plateThe depression was visible even before V. Pârvan started the excavations in Histria [19] - that mean before "cleaning" the late Roman walls mentioned to have existed here. The sections made in 1944 demonstrated that the actual soil is very close to that of the cliff. This situation imposed (as seen in the Domus area) massive demolition in the moment of construction of the Roman monuments of the area. The same sections showed that in the civilian basilica all the underlying levels were destroyed by the Roman construction. The foundations of the basilica are descending to the cliff, that was levelled.

[19] V.Pârvan, in Boabe de grâu, p. ; Buletinul Comisiunii Monumentelor Istorice, p.

"B" AREA (north-east)

There is no available data for this area because of the very few excavations made here only before the World War II. In conclusion, we have tried to prove that a new approach of an old problem -- the Histrian topography -- can bring new and maybe unexpected solutions. We hope that the fact the solutions are rising more new questions will not be considered as guilt of the author.