REVISTA DOCTORANZILOR ÎN ISTORIE VECHE ŞI
Transcription
REVISTA DOCTORANZILOR ÎN ISTORIE VECHE ŞI
REVISTA DOCTORANZILOR ÎN ISTORIE VECHE ŞI ARHEOLOGIE ReDIVA THE POSTGRADUATE JOURNAL OF ANCIENT HISTORY AND ARCHAEOLOGY II/2014 CLUJ-NAPOCA 2014 3 CONTENTS STUDIES Mariana Prociuc, Vlad Codrea Archaeozoology and palaeontology of the Subpiatră Cave (Bihor County, Romania) 11 Aurora Peţan An unknown stone structure in Sarmizegetusa Regia’s sacred zone recorded in writings of the 19th century 28 Mátyás Bajusz, Aurora Peţan Two bronze bracelets with looped and twisted ends from the notes of Téglás István 41 Csaba Szabó Discovering the gods in Apulum: historiography and new perspectives 53 Radu Iustinian Zăgreanu, Claudiu Ionuţ Iov A Roman funerary stela from Porolissum 83 Alexandra Teodor The roman defensive system(s) of Tomis. Some issues in the light of the current knowledge 92 REVIEWS Todd L. VanPool, Robert D. Leonard, Quantitative Analysis in Archaeology, Wiley-Blackwell, 2010. (Laura-Simona Draşovean) 149 Luca-Paul Pupeză, Veacul întunecat al Daciei, Cluj-Napoca, 2012. (Raluca-Eliza Bătrînoiu) 153 Ioan Piso, Viorica Rusu-Bolindeţ, Rada Varga, Silvia Mustaţă, Ligia Ruscu (eds.), Scripta Classica. Radu Ardevan sexagenario dedicata, Cluj-Napoca, 2011. (Aurora Peţan) 159 Rada Varga, The Peregrini of Roman Dacia (106-212), Cluj-Napoca, 2014. (Cosmin Coatu) 167 7 THE ROMAN DEFENSIVE SYSTEMS OF TOMIS Some issues in the light of the current knowledge Alexandra TEODOR PhD Candidate, “Ion Mincu” University of Architecture and Urbanism, Bucharest, RO E-mail: [email protected] Abstract: The Roman fortifications of Tomis (today, in the city of Constanța, Romania) have entered in the attention of our scientific circles about a century ago, with the discovery of an urban defensive wall segment and a tower. Though some remarcable discoveries have followed since, no thorough study integrating all known data regarding the city’s defensive elements and their topography has been published so far. This paper’s aim is not to repair this at once, but to sketch some methodological approaches towards integrating available data, to (re)activate some issues regarding its general chronological outlines and to draw some further research directions and possible conclusions that came out of this approach. Using digital technologies, a large amount of the available data directly and indirectly regarding the urban defensive systems of Tomis were integrated. Architectural more or less detailed plans (the archaeological evidence), textual informations (descriptions) and funerary evidence (spatial and chronological coordinates) were assembled on the same topographical support, allowing thus a multilayered analysis. Theoretical constructions, such as the spatial and temporal relationships between funerary areas – defensive systems – roads – city gates – streets, starting from, but also complementary to the archaeological evidence, were used in order to provide new arguments for some older hypotheses, to debate others or to emphasize new ones. The main outcomes of this study are a different perspective – though not entirely new – than the ‘consecrated’ theses concerning the defensive systems (alongside the implicit effects on reevaluating the city’s urban evolution, not detailed here) and, as an indirect result, new clues regarding a possible partial continuity of the Roman urban street network up to our days. Keywords: Roman fortifications, Tomis, data integration, urban archaeology, antique urban topography. ReDIVA II/2014, p. 92-146 The roman defensive systems of Tomis 1. Introduction Tomis’ antique history1 can be seen through two rather paradoxical attitudes. On a side, its fame and considerable size2 have determined a wide scholar’s enthusiasm which at its turn generated a huge bibliography3. On the another side there’s a large gap in awareness, knowledge and interest. The first is represented by devoted or occasional researchers of the antique city, and the second by today’s most of local officials, citizens and tourists. Between the two sides there are various attempts to promote history, unfortunately not all updated or adequate to the current state of knowledge4 (Figs. 1 and 2). A briefing note for those unfamiliar with the site’s general coordinates: Tomis was, throughout Antiquity, one of the largest cities on the western coast of the Black Sea, founded by Greek Milesian colonists sometimes in the 6th Century BCE; in the Early Roman Era it became at some point the Metropolis of the Greek cities’ association of the western Pontus Euxinos (the Black Sea); in the Late Roman era, Tomis was the Capital of the Roman Empire’s easternmost Danubian border province, Scythia. Today, the former antique city is overlapped by the modern Constanța’s city center, in the area known as “Peninsula”; surprisingly or not, the city is still a great commercial hub on the Sea’s shore. For details on Tomis’ history in a recent historical and archaeological synthesis, see Buzoianu, Bărbulescu 2012, p. 127-40 (Greek and Hellenistic), p. 140-78 (Early Roman), p. 178-212 (Late Roman), and p. 213-7 (post Roman), with further literature. 2 Tomis was the largest antique city of Dobruja historical region, reaching more than the 60-70 ha we can measure today, as the city may have lost important land parts to the Sea, see Scarlat 1975. 3 See Buzoianu, Bărbulescu 2012, p. 11, note 1 for the most important titles regarding the city’s history. 4 The most recent and fashionable are the 3D reconstructions, usually available online, see for example catheofilm.ro (“Reconstrucție 3D – Termele Romane Tomis” [The Roman Baths], which barely has anything to do with the archaeological and topographical reality, and “Reconstrucție 3D – Portul Antic Tomis” [The Roman Harbor]), cerva.ro/tomis/ (see also Popovici et al. 2010) and cjc.ro. In the last two cases – rather worthy of our attention –, despite their declared historical and archaeological assistance and their stated objectives (see ‘About’/‘Echipa’, ‘About’/‘Descriere’ and ‘Obiective’ sections on their websites), they both reflect too little attention to topographical and/or historical and archaeological reality. The question one needs to answer is whether we truly want the public to go from barely knowing anything to falsely knowing it all, rather than adequately informing them about the real level of knowledge and understanding. For instance, none of these reconstructions represents one of the most impressive construction works of Tomis – the water gallery system (see below, 4.3.2), as an important ‘layer’ of the antique city. 1 93 Alexandra TEODOR The antique city’s historiography actually has only a few dedicated thematic monographs5 besides the general ones (see notes above), and among them even fewer on the antique urban structures or related (such as necropolises, see below, 2 and 4.1). We may also say that at least in what concerns the urban topography issue, quite rarely the authors’ critically deal with already published and interpreted data6 and, when necessary and possible, offer a new interpretation7. Furthermore, so far there was apparently no attempt to integrate available data into an adequate topographical support8,9. Papuc 2005, a monograph dedicated to the water systems of Tomis (most of it a synthesis of previous studies on the subject), with an useful historical and archaeological introduction; Stoian 1962, on epigraphic evidence available at that time, also with an useful and detailed introduction; Canarache et al. 1963, dealing with the antique sculptures of Tomis, completed recently by other studies, such as Covacef Z 2011 (though not exclusively on Tomis). See this small inventory compared the ones of sites under systematic archaeological research, such as Histria (with its 14th recently published volume in the dedicated monograph series), to remain only in the Romanian field. 6 Toma 2010, p. 59 (on named here Line E?’s problematic archaeological argumentation, see also below, 4.2.1.D). 7 Toma 2010, p.64-8, 71-72; Suceveanu 1969, p. 348, 349, 351-2 (both on the consecrated dating of the so-called Late Roman city walls and of the so-called “Mosaic Edifice”). 8 By this I mean data integration using dedicated digital technologies. Natalia Toma (2010) made an attempt by simply vectorizing three very important old city plans (Toma 2010, p. 58, figs. 1 and 2, 62, fig. 3) and approximately integrating them into a topographical restitution of the antique city (Toma 2010, p. 62, fig. 3); the result is reliable, but not precise. The only – but large – issue with the above mentioned integration method is that especially old plans frequently have proportion and deformation errors compared to the actual precise topographical or orthophotographic representations (see comments at Toma 2010, p. 57 and Georgescu 1998, p. 55-6, referring to general problems related to the topographic archaeological documentation). This may not mean much when largely discussing urbanistical issues, but it becomes a crucial problem when discussing topographical relations between a defensive line’s segments on a detailed scale; this is why the author couldn’t realize, for instance, the problems related to defensive lines named here A and B, see below, 4.2.4. Therefore, “dedicated” digital technologies would mean something more than simply and approximately merging data from different planimetric representations; it means, however, confronting them with a reliable topographic support and further operating depending on their fair or unfair correspondence: either direct vectorization or vectorization preceded by georeferencing (see for example Polonic’s plan, here at Fig. 5). 9 This study and the precedent one, Teodor 2012, are the result of such intentions. A more extended file for Tomis’ urban configuration is prepared for 5 94 The roman defensive systems of Tomis The bibliography on Tomis’ fortifications10 makes no exception to the mentioned state of art11. As a matter a fact, a closer look reveals some issues left unsolved and/or ignored maybe for too long (see discussions below, especially section 5). It’s a known fact that in Antiquity, and for a long time after, the fortifications were one of the urban structure’s main components. It should also be a known fact that if we fail to understand the urban fortifications’ character and their evolution, we fail to understand the whole city. This presentation is meant to direct the reader12 towards a refreshed and hopefully better understanding of the archaeological evidence regarding the antique city’s fortifications, with a special look on their topography. Last but not least, it is to be noted that the antique city’s overlapping by the modern city of Constanţa is one of the biggest impediments in documenting its physical configuration; just recently the greatest opportunity for collecting some more information was successfully lost13, meaning that for some while – which means until other large investments will come up in the old city center – we can only go further in our research with what we already have. the PhD dissertation regarding late antique urbanism in Scythia, which is to be defended, hopefully, towards the end of this year. Most of the issues discussed here will also be presented there, in Romanian, the official language of the doctoral programme. 10 Cliante 2006; Papuc, Lungu 1998; Georgescu 1998, p. 60 sqq., 78-90, 158-64; Rădulescu 1995-1996; Rădulescu 1991, p. 23-9; Bănică 1991; Coja 1982, p. 101 and fig. 221; Cheluţă-Georgescu 1977, p. 258-9 and pl. II; Stoian 1954; Canarache 1961, p. 16-21 and fig. 1, Lambrino 1936; Pârvan 1915; Polonic mss., f. 51-62 and city’s plan; except for the titles where mentioned (not exclusively dedicated to Tomis’ fortifications), see also plans and pictures which are part of them. 11 See a former experience with Tomis’ fortification systems, Teodor 2012, p. 4-5 (2.2.2) and fig. 2, with comments regarding different plans representing the same structures. 12 I mainly address the ones already familiar with the site’s features, but it could also be an useful look over for those wanting to get familiar to it. 13 Latest infrastructure investments in Constanța’s historical center completely ignored, against all laws, the archaeological layers’ proper documentation. See a few opportunities described here (for now only in the local press): gzt.ro-1 2013 and gzt.ro-2 2013. 95 Alexandra TEODOR 2. Methodology and Aims The study is based on published or other available14 data, without pretended exhaustivity, and on personal topographic observations relying both on published planimetric representations and other relevant localizable data, all mapped on a georeferenced topographic base15. Urban defensive structures can be referred to, even though indirectly, in the most various sources. Therefore, after a briefing on the generally accepted scenario regarding known fortifications and their chronology (3, The Consecrated Story), a larger set of relevant identified data on the issue will be listed (4, Available Data): (4.1) the known necropolises’ spatial distribution as main topographical framing for the residential area; (4.2) the archaeological evidence on (4.2.1) uncertain and (4.2.2) certain fortification lines16 with their known elements’ typology and construction system, as well as (4.2.3) topographical issues regarding compatibility to one another; (4.3) the complementary evidence, comprising (4.3.1) the known stratigraphic data, (4.3.2) related urban infrastructure and the (4.3.3) numismatic iconographical evidence. After parsing these, one might realize that some of the so far conveyed accounts regarding this antique city’s urban fortifications reveal as overrated, unreliable hypotheses, constantly propagated through its historiography, although at some times few voices questioned them, see (5) Discussions and Hypotheses. (6) Further Research directions, meant to complement this preliminary study’s directions, will be sketched before (7) Conclusions. Pamfil Polonic’s manuscript (Polonic mss.); the extremely useful urbanistical documentation (PUZ 2012) and the orthophotographic support (ANCPI 2005, 2008). 15 Topographic data was assembled in AutoCAD 2010 (Educational version) in Stereo 70 (the Romanian standard topographical projection system; in the area discussed, it has ca. 2.50 declination to the West compared to the geographical North). All plans were scanned to avoid distortions as much as possible (except Pârvan 1915, p. 440, fig. 4, photography, but errors are not substantial compared to the orthophoto, see here Fig. 9a). 16 They’ll be presented as distinct lines (defensive Line A, B, ..., F?) only for methodological purposes, since their spatial and chronological relations are exactly the ones being questioned here. 14 96 The roman defensive systems of Tomis Besides the archaeological evidence, theoretical concepts will also be used in the methodological approach. The high relevance of funerary data17 is underlined here as a possible starting point related to residential area (fortified or not), but also on two other crucial and highly connected urban elements: roads/streets and city gates. Although quite trivial arguments, they’re worth mentioning: (arg. 1) for the Greek and Roman Eras, generally no burials were allowed inside city walls18,19; (arg. 2) for the same period, generally, streets outside cities (roads) were flanked by funerary mounds (by reverse, it means mounds usually indicate roads)20; (arg. 3, starting from arg. 1 and 2) the limits of the funerary area towards the city indicate the utmost spot to find the gate on a route track, i. e. the gate can be found anywhere behind that spot. All the mentioned elements – roads (Road A, B etc.), funerary/ residential limits (NL1, NL2 etc.), gates (G1, G2 etc.) and street (Street 1A, Street 1B) – will be first pointed out as the above mentioned theoretical concepts (therefore theoretical points and lines). They We have to note here that there is no updated monograph on funeral evidence at Tomis, for any of the Greek or Roman periods; the only existent monograph on such themes is Bucovală 1967 (Hellenistic period). Most important for our purposes, there is no updated funeral evidence chart available; integrating separately charted evidence in various publications and identifying the spatial location of simply described points, with no cartographic support, can be really time consuming. Therefore, data presented here is not pretended to be exhaustive. 18 The known examples of funerary complexes inside the city walls in the Greek world, although very rare, regard the so-called heroes’ tombs, see Rykwert 1995, p. 33-7, especially 33 and notes 41-42, 35 and note 57; for the Late Roman world, at least in Scythia, intra muros burials are only exceptional, see Soficaru 2011, p. 137-9 with bibliography, 138 Table12, 178-9, Table 17. 19 That is why, when speaking about urban fortification lines, the funerary evidence is the one we shall call, metaphorically, “the negative evidence”. Therefore, when we have no archaeological data (“the positive evidence”) on the fortifications and/or extra muros residential area, and/or their chronology is doubtful, one of the main data we can rely on is the funerary evidence, because it indicates the maximum possible spread of the fortifications/residential area in the correspondent chronological frame; also, no matter how ‘well known’ they both are, it is always useful (and also safer) to confront their chronology. 20 For Tomis, see Bucovală, Pașca 1992, p. 271; for other references, see also Constantin et al. 2007, p. 263-4 and n. 47. 17 97 Alexandra TEODOR must not be confused with the real, archaeologically attested elements (where it is the case), which will be generally approached afterwards, such as certain defensive lines (Line A, B etc.), certain gates (Gate 1, Gate 2 etc.), or simply hypothetical defensive lines, based on different arguments (Line E?, Line F? etc.; Gate 4?, Gate 5? – all pointed out with ‘?’). The scope of this bionomical system (theoretical constructions/concepts – real elements) is to provide hints regarding the possible urban configuration and evolution, by comparing and integrating the two parts, as we shall see for instance in section 5.5. 3. The Consecrated Story21 (see Fig. 3) 3.1. It is known that not only the Roman city had its urban fortification system, but also the Greek one from which it subsequently developed. Not much is known about it though and the only reliable hints regarding its localization are given by the funerary evidence (see below, 4.1). It is generally believed to be placed near today’s Ovidiu Square. For the record, so far no published archaeological investigation has documented any Greek fortification in Tomis. 3.2. Concerning the Roman fortifications, the story goes like this: an Early Roman fortification most probably existed, somewhere between the Greek (unknown, see above) and the (consecrated) Late Roman ones; the statement is based on some older funerary evidence and related hypotheses, somehow “confirmed” by time-to-time discoveries in the old city center (see also below, 4.2.1.B and C). To sum up, there is no hard evidence and no archaeological record regarding this Early Roman precinct. 3.3. Last but not least, there is the so-called Late Roman fortification. One of its supposed first segments was discovered at the beginning of the XXth Century and since then many others. We can count now few curtain segments, two gates with rectangular bastion towers, one gate with long U-shaped towers, one short U-shaped tower and at least one rectangular tower. Its first phase This is only a briefing. For details and references for all data mentioned, see the next section, 4. 21 98 The roman defensive systems of Tomis (named here Line A) is generally dated at the end of the third century AD, with few important alterations and/or additions – one in the end of third and first half of the fourth century AD (named here Line B), and others in the sixth century AD, including a so-called extension (named here Line C); for all, see below, 4.2.1.D and 4.2.2. 3.4. On a secondary plan, there is a general hesitating position about the presence of a defensive line on the Sea shore, meaning whether the city would have had either an open defensive line, or a closed one. Thus, the story regarding the antique Tomis ends by concluding that the city’s evolution reached its peak in the Late Roman era22, when it played the important role as the capital of Scythia, despite the barbarian attacks that seriously and repeatedly damaged the city in this period. 4. Available Data 4.1. Funerary evidence (see plan at Fig. 4) 4.1.A. Only a few Classical and more Hellenistic graves have been found so far23. Their various density and spatial distribution allows us some observations. First, one could quite clearly delimitate the residential area from the funerary one for this period; in this matter, one thing to know for sure is that the fortified city’s limit was smaller than the supposed Late Roman one (see also Fig. 3, Line A, for comparison). Another hint regards two possible roads, along the areas where the funerary evidence is more dense24: Zahariade 2006, p. 75; Rădulescu 1995-1996, p. 92. CCA 2008 (Mihai Viteazu Street, southern segment); Bucovală 1995, p. 73-74 and pl. I (Mircea cel Bătrân – Decebal Street crossing); Chera-Mărgineanu 1979, p. 250 (Cuza Vodă Street, eastern end), passim; Bucovală 1975, especially p. 375, 387 (Tomis Bd. – Negru Vodă Street crossing); Bucovală 1968, p. 269-70 and fig. 1; Barbu 1961, p. 204-5 and fig. 1 (Ferdinand Bd. – Tomis Bd. – Negru Vodă Street crossing area). 24 For brevity, the road’s routes presented here and in the next sections (see the plan at Fig. 4) were actually correlated to the already known possible antique routes – Polonic mss., f. 56-8; see also Toma 2010, p. 61-3 (gates), 68-9 (roads) and 58, 62, figs. 1-4. Therefore, we might say the roads traces 22 23 99 Alexandra TEODOR Road A, towards northwest (supposedly heading to Histria), and Road B, towards southwest (heading to Callatis), most probably the main roads25 connecting the city to the proximate territory and the main regional centers in the Greek Era. Both roads are also confirmed by later funerary evidence as main routes, meaning they were probably used quite constantly throughout all Antiquity. 4.1.B. Early Roman funerary evidence dated in 1st-2nd Centuries CE26 is quite compact in space, but that’s probably due to its relative scarcity; still, it is very relevant from a topographical point of view, because its distribution enters deeply in the Peninsula. Threfore, the southernmost funerary evidence indicates with high probability the funerary/residential limit (named on the plan NL1) and, by consequence, the utmost position of the city gate in this phase (G127), most probably located on Road A’s intersection with the city wall; this theoretical point is one to remember. 4.1.C. Early roman evidence dated in 2nd-3rd Centuries CE, quite consistent28, spreads along already existant roads A and B and didn’t only result from funerary evidence, since their spatial distribution can be confusing or insufficient sometimes, but rather we relied on funerary evidence for supporting the roads’ possible dating and evolution. This is why – but also because general topography implies it –, for instance, Road B doesn’t have a straight course, as a strictly hypothetical one would have. 25 Related to roads directions leaving out of Tomis, see Georgescu 1998, p. 93; Canarache 1961, p. 20. 26 gzt.ro-1 2013 (first half of the 1st Century CE, next to Hunchiar Mosque); CCA 2008 (Mihai Viteazu Street, southern segment, 1st Century CE); Georgescu 1998, p. 71 (imprecise, northern area of the Peninsula’s; end of 1st Century CEfirst half of 2nd Century CE; possibly referring to the same funerary evidence as Rădulescu 1991, see below in this note); Bucovală, Pașca 1988-1989, p. 160 (“La serpentină” area, close to modern Harbor Gate 3, 2nd Century CE); Bucovală 1968, p. 269-70 and fig. 1, 303 (I.G. Duca Street); Rădulescu 1991, p. 35-6 (south to Gate 3, Mircea cel Bătrân Street – Ibis Hotel area, 2nd Century CE); Zirra, Alexandrescu 1957, p. 89 (Marc Aureliu Street, 2nd Century CE). 27 Neither this, nor the next ones are to be considered properly gate positions, for which see below, 5.5. 28 Only the most relevant titles: Băjenaru, Dobrinescu 2008, p. 190, 199, pl. 1 (George Enescu Street); Bucovală, Pașca 1992, p. 241-2, fig. 1 (I.G. DucaConstantin Brătescu crossing, in the middle-western area); Bucovală and Pașca 1991, p. 185, 186, plan 1 (the southwestern remotest area); Barbu 1961, p. 205, 207 and fig. 1 (largely both southwest and north areas). 100 The roman defensive systems of Tomis also slightly suggests one more, heading west, named Road C – that may have already existed in the previous phase –, but also another, parallel and west to Road A, named here Road D (for which archaeological funerary evidence is completed by other sources, see below, 4.1.E;). We can trace two relevant limits on this funerary area: NL2, towards north-northwest, and NL3, towards west. It is notable these limits haven’t changed much afterwards, as the evidence points out (see the plan and below, 4.1.D). Having this in consideration, it is highly probable that the space between NL1 (see above, 4.1.B) and NL2-NL3 was now affected to the city (intra or extra muros, yet uncertain); we also assume that in these times a new fortification has been built (see below 4.2.1.C). Therefore, somewhere behind the intersection of the identified roads and the known limits of the funerary area, one should find, theoretically, the city gates; for methodological purposes, the utmost gates’ positions (G2-G5) are marked on the plan (see also below, section 5.5, which of them could have actually functioned). 4.1.D. All Late Roman funerary evidence, dated throughout 4th-7th Centuries CE29 is distributed along the same roads mentioned above. Also, their limits towards the city are not very different compared to the 2nd-3rd Centuries’ evidence; from a topographic point of view, the main difference noted so far is their larger spread between the existent roads and towards the territory. 4.1.E. The mounds noted by Pamfil Polonic on his city sketches (Fig. 5), drawed towards the end of the XIXth Century, could be considered another funerary evidence. Of course, the structures cannot be accurately dated, but we can assume that they were originally pagan funerary structures and that they were used until later times, since they were best visible until recently; a plausible Only the most relevant titles: Bucovală, Pașca 1992, p. 241-2, fig. 1 (I.G. Duca-Constantin Brătescu crossing, in the middle-western area); Bucovală, Pașca 1988-1989, p. 123, 124 with map, 160 (with a 4-5th Century collective grave problematic for its imprecise location, possibly close to the city walls); Chera-Mărgineanu, Lungu 1983, p. 217, 228-230 and 229, plan (north, west and southwest areas in the city’s proximity); Chera-Mărgineanu 1979, p. 247-50 (I.G. Duca-Cuza Vodă crossing and remotest areas); CheluţăGeorgescu 1977, p. passim and pl II (distributed in all directions); Barbu 1961, p. 205 and fig. 1 (largely eastern shore and proximity). 29 101 Alexandra TEODOR dating would be, therefore, 2nd-3rd Centuries CE and/or before. What they clearly indicate – and Polonic also points out – is another ‘main’ road, called here Road D (also indicated by certain funerary evidence, see above, 4.1.E); it should be, therefore, contemporary to Road C, or earlier. The utmost position of the possible gate corresponding to this road is G3, noted on the plan. 4.2. Archaeological evidence 4.2.1. Uncertain defensive lines (see again Fig. 3) 4.2.1.A. There is no archaeological (documented and published) evidence regarding the Greek and Hellenistic Era’s urban fortification. It is generally assumed that it was much smaller than the so-called Late Roman one and its location is now thought to be somewhere north of today’s Ovidiu Square30. Based on the funerary evidence, theoretically, its limit could reach further, closer to the southernmost Hellenistic grave31; still, being given the quite modest evidence (as quantity), this hypothesis is yet premature. 4.2.1.B. From the very Early Roman period of Tomis (1st Century) our only direct evidence so far are Ovidius’ statements regarding the city walls; he mentioned their small length and modest defensive qualities32. The funerary evidence (see above, 4.1.B) indicates its narrow layout and withal the possible defensive limit for largely the same period (about 1st Century-first three qarters of the 2nd Century), Rădulescu 1995-1996, p. 91. Recent infrastructure works in the named spot revealed a large stone structure seemingly to be the Greek precinct, according to the archaeologist who observed them; unfortunately, archaeological investigation wasn’t part of the investors’ programme (gzt.ro-1 2013), therefore, this visual observation can only be taken, at least for a while, as a highly probable hypothesis. Other indirect clues come from different inscriptions mentioning city’s guards, basically indicating the same characteristics regarding the defensive line’s length as being rather short, see Stoian 1954, p. 558; see also Georgescu 1998, p. 59-65, for known hypotheses and his own, referring to an area smaller than the one delimited here by funerary evidence, but insufficiently argued. 31 See also below, 4.2.1.B and notes. 32 Georgescu 1998, p. 69-70; Rădulescu 1995-1996, p. 91. See also Covacef P 1998, p. 261. 30 102 The roman defensive systems of Tomis named here Line G?. The hypothesis that this epoch’s defensive line could have been one and the same with the older Hellenistic/Greek one has been also stated.33 4.2.1.C. From the later part of the Early Roman period, the Severan dynasty Era is known as a prosperous one, and scholars generally believe it had to have been marked by an extension of the formerly (more) modest town. The largely accepted hypothesis, although based on scarce arguments (see below, 4.2.1C), is that the defensive line of this period, named here Line F?, is to be placed somewhere between the earlier fortification (Line G?, see above, 4.2.1.B) and the so-called Late Roman fortification (defensive lines A-B-C, see below, 4.2.2). Its southwestern end is considered to be a defensive element that was never properly documented – therefore it is uncertainly dated –, named here Tower 5 (see also below, 4.2.2, where it is associated to a different defensive line, Line D)34. Line F?’s direction is approximately parallel to the common directions of defensive lines A and C, crossing the peninsula eastwards compared to them. For the general line the assumed arguments are grounded on unpublished archaeological investigations, only briefly described, without the necessary archaeological documentation attached35. Still, although ungrounded in our opinion, this hypothesis cannot yet be completely rejected, until some direct or indirect archaeological arguments will be prompted. 4.2.1.D. For the Late Roman Era (late 3rd-early 7th Centuries), the ‘consecrated’ story is that defensive Line A (see below, 4.2.2), dated at the beginning of this epoch and considered to having functioned Georgescu 1998, p. 71, but see also the section above, with notes. Rădulescu 1995-1996, p. 91; it is very likely that this tower was only considered by the author based on its spatial proximity, which is definitely not an argument. 35 Rădulescu 1995-1996, p. 91; Georgescu 1998, p. 74-6, describes a 2.6 thick two-faced wall with a Hellenistic-like parament; apparently, there is no reason why this couldn’t be the Hellenistic city wall, since it is also compatible to the Hellenistic known funerary area. This hypothesis actually circulated for a while, see Coja 1982, p. 101; Barbu 1961, p. 204. By any means, this would imply that there was also a narrower city wall for the first part of the Early Roman Era, also proved by contemporary funerary evidence – a possibility which maybe should be explored from an historical point of view. 33 34 103 Alexandra TEODOR throughout all of it, was until the 6th Century closed westwards by a presumed defensive Line E?, which was then abandoned and replaced by Line C36 (for which see below, 4.2.2). This Line E? is figured on most of the city’s plans representing the defensive system as a rather certain defensive line 37 and its existence is practically considered as a sure fact, although the hypothesis was never archaeologically proved ever since its formulation, and its grounds are rather shaky38. 4.2.1.E. No thorough investigation ever concerned the hypothetical shore defensive line, therefore references to it are rather hesitating39. Two possible segments of a defensive line were noticed more than a century ago close to the shore, on the southern and southeastern sides of the Peninsula, see Line H?; since no precise topographical documentation was published for these structures, their location and orientation were approximated, having as landmarks some buildings 36 Cliante 2006, p. 249-50; Georgescu 1998, p. 163-4; Rădulescu 1995-1996, p. 86-88; Cheluţă-Georgescu 1977, p. 259 and pl. II; (see also next note). 37 Băjenaru, Dobrinescu 2008, p. 199, pl. 1; Cliante 2006, p. 253, fig. 1; Lungu 2000, plan at page 129, with legend; Rădulescu 1995-1996, p. 90, fig. 5; CheluţăGeorgescu 1977, p. pl. II. On the other hand, Papuc, Lungu 1998 didn’t seem to agree, but this is only indicated on the plan (or it is simply an omission), see p. 202, fig. 1, with legend; similar at Toma 2010, p. 62, fig. 4, ignoring the hypothetical line. 38 All the references to the city’s extension from the 6th Century quoted above automatically imply this defensive line’s presence. Its hypothetical existence was first formulated by Cheluţă-Georgescu 1977, p. 258: “(...) la 35 m sud de această intersecție [Ferdinand Bd.-Mihai Viteazu], a apărut (la -4,50 [m]) o îngrămădire masivă de blocuri și pietre prăbușite, care făceau parte dintr-un zid masiv, orientat S.V., spre actualul Muzeu al Marinei (...)” (basically, the supposed defensive walls’ presence was assumed based on a “(...) large stack of fallen stones as part of a massive wall oriented southwest, towards Marine Museum”, approximate translation; see the plan at Fig. 3). In order to explain both the massive wall’s presence there and the already known segments of what we call here Line C (see below, 4.2.2), the same author issued the 6th Century city’s extension hypothesis, see p. 259 and pl. II (and also above note). These two correlated hypotheses were also propagated in the more general historical literature, see Zahariade 2006, p. 75. Toma 2010, p. 59 and 62, fig. 4, rejected Line E?’s theory invoking insufficient arguments, but didn’t go further, declining the correlated ‘extension’ hypothesis. 39 Rădulescu 1995-1996, p. 91 and note 29. 104 The roman defensive systems of Tomis from the XIXth Century with known location. They most probably belong to a single defensive system, a continuous line along the high shore of the Peninsula, possibly as well as Line D (at least from some point in time, see also below, 4.2.2). Another tower was mentioned in more recent times, close to the southern segment of Line H?.40 It’s worth mentioning that the same construction system as Gate 1 of Line A (opus quadratum or similar, dry assemblage, see below, 4.2.2) is attested by older references for the defensive Line H?.41 4.2.2. Certain defensive lines42 (see again Fig. 3) There are known and archaeologically attested four defensive lines so far – lines A, B, C and D. Segments of line H? (presented above, see 4.2.1) have been seen as ruins some time ago and in the meantime have disappeared completely or were partially preserved under the modern city. Line A (see also Fig. 6), closing the city towards northwest, consists of ca. 225 meters of curtain (ca. 3 meters width43) fitted with two similar gates (Gate 1 and Gate 2) and one short U-shaped tower (so-called The Butchers’ Tower, named here Tower 1). Except for the later element, dated in the 6th century based on epigraphic evidence (well, at least its reconstruction)44, the defensive line’s consecrated dating is at the end of the 3rd Century, based exclusively on some 40 Rădulescu 1995-1996, p. 91 (tower on the shore, at the southern end of Arhiepiscopiei Street, see the plan at Fig. 3). For the older references, Polonic mss., f. 53-5 (curtain and two inside towers next to the old Casino, situated on the same approximate location as today’s Casino, and a curtain with a tower next to Carol Hotel; see the plan at Fig. 3); Michel 1862, p. 10. 41 Polonic mss., f. 55, Michel 1862, p. 10. 42 This short presentation isn’t meant to completely describe the defensive elements’ architecture (documentation to be found in the literature quoted below), but to point out the relevant issues in our discussion. 43 Canarache 1961, p. 19. 44 Rădulescu 1991, p. 23; Pârvan 1915, p. 416 sqq. and 418, fig. 1 (plan), 438, fig. 3 (elevation). An inscription was found in the tower’s exterior parament, saying the Butchers’ association funded its building (Pârvan 1915, p. 419-20, 437-9; Rădulescu 1995-1996, p. 86); still, it could be a rebuilding, but this now quite difficult to establish since its structures were mostly ‘restaurated’ (Rădulescu 1995-1996, p. 83). 105 Alexandra TEODOR coins found in curtains’ emplecton, all dated between the times of Aurelianus (270-275) and Probus (276-282)45. Although at first opinions were more reserved, admitting a simple reconstruction dated in these emperors’ times or later46, quite unexplainable most of later opinions focused exclusively on considering this period as the definite time of construction47 (see also below, 4.2.3). Line A, Gate 1 (also named, improperly, West Gate; see Fig. 9a, left). The most relevant documentation for its architecture is Teodor Bănică’s study48 – with details (construction techniques, at least) which, by obvious analogy, can be largely considered also for Gate 2; unfortunately, this study was largely ignored in the further literature49, although it points out some very relevant architectural discrepancies that most probably indicate some larger chronological gaps between Line A and Line B (see below) than generally considered50. To sum up, the two gates’ towers have a clearly distinct architecture than all rest of the towers found (and more properly documented) at Tomis so far. They were probably not towers, technically speaking, but bastions, named here TB1 and TB2 for Gate 1, respectively TB3 and TB4 for Gate 2; their dimensions are quite similar, the fronts measuring 5 up to 6 meters, and their sides ca. 2.8-3.5 meters. Distances between bastion-towers measure ca. 5.3 meters for Gate 1 and ca. 4.5 meters for Gate 2. They were built in opus quadratum, with no lime, most probably using metal clamps for assembling51. Walls’ thickness was small, of only one stone block, Cliante 2006, p. 249-50; Rădulescu 1991, p. 28; Georgescu 1998, p. 77-79, 89, 162; Bănică 1991, p. 62 (though not convinced, see also the final phrase at p. 65; Canarache 1961, p. 16-7; Barbu 1961, p. 204. 46 Canarache 1961, p. 16-7. 47 With only few exceptions see Suceveanu 1969, p. 348 sqq. and, largely based on the first, Toma 2010, passim. 48 Bănică 1991. 49 Not mentioned by Rădulescu 1995-1996 (a synthesis study on the city’s fortifications issue); Papuc, Lungu 1998; Cliante 2006 and Toma 2010. 50 Same kind of observations personally documented and stated by Toma 2010, p. 60. 51 According to Biernacka-Lubańska 1982, p. 213, this system is typical for the period between 2nd century-middle of the 3rd Century in Moesia Inferior (from which Tomis was part of at that time) and Thracia; see more details for the construction characteristics of the urban fortification of the period at p. 212-214. 45 106 The roman defensive systems of Tomis of which length could have reached 1.5 meters. The towers were most probably filled with earth, since no emplecton was found on spot, at the discovery time. The exterior facing of the towers and the connected curtains still preserved a special treatment (embossment) of the stone blocks which clearly suggests that the material was specially prepared for this construction. On site analysis is now partially compromised by subsequent ‘restoration’ works, therefore Bănică’s study, based on unpublished documentation prepared before the ‘restoration’, is very important. Line A, Tower 1 (The Butchers’ Tower; see Fig. 9a, right). It is an U-shaped tower, measuring 7.5 meters wide and 9 meters length (both exterior); the walls’ thickness is about 2.5 meters. The walls were built in opus implectum faced with opus incertum – larger stone on the outside, small stone on the inside.52 Line A, Tower 253 (east to Tower 1, west to Gate 3; see again Fig. 6), rectangular plan, temporarily came to light in the 1960’s, but it was never properly published. Its only (somehow) detailed and precise topographical representation was to be found in some recent urbanism documentation for the city center; the tower’s localization seems to correspond to the one indicated by Canarache. For all possibilities, this tower must have been – at least for some while – a part of the defensive Line A, as all plans that take it into account also seem to indicate. Only approximate dimensions are known for this projecting rectangular tower, of about 8 meters exterior depth, 10 meters length (front) and ca. 3 meters walls’ thickness54. Compared to TB1-4, it is most probably a true tower and not a (filled) bastion Pârvan 1915, especially p. 416-7 and 439-41; 418, fig. 1, 438, fig. 3, 440, fig. 4. Canarache 1961, p. 20; PUZ 2012. The tower is represented on various plans (approximately) mapping the city walls at Zahariade 2006, p. 68, fig. 27 and 71, fig. 29, Rădulescu 1995-1996, p. 253, fig. 1; Chera-Mărgineanu, Lungu 1983, p. 229, pl. 3; Coja 1982, p. fig. 221; Cheluţă-Georgescu 1977, p. pl. 2; Bucovală 1968, p. 270, fig. 1; Barbu 1961, p. fig. 1 (where the tower was misplaced); Canarache 1961, p. pl. 1. Curiously, at some authors, the tower was not represented after Gate 3 was rediscovered and its location and orientation were better determined (1998), see Papuc, Lungu 1998, p. 202, fig. 1 (although its presence was known and discussed, see p. 207); Cliante 2006, p. 253, fig. 1; Băjenaru, Dobrinescu 2008, p. 199, pl. 1. 54 According to PUZ 2012 documentation. 52 53 107 Alexandra TEODOR tower, therefore its construction date could be later than the gates’ 1 and 2. From Line B (see also Fig. 7) it was yet uncovered and published only a gate ensemble (Gate 3), containing two elongated U-shaped towers (named here Tower 3 and Tower 4, see Fig. 9b) flanking the gate and the adjoining curtains55. It is to note that the gate’s baseline’ direction is practically one and the same with the modern Street Dragoş Vodă’s axis56, and also quite different to Line A’s direction (see below, 4.2.3). The U-shaped towers are one of the biggest known of this type in Scythia – if not the biggest57. They measure about 11.2 x 16.8 meters (T3), respectively ca. 11 x 16,4 (T4), with wall thickness at foundation level of about 3.7-4 meters, but probably about 3 meters at the elevation level58. An interesting functional detail is to be mentioned: on the western wall of the Tower 3, a poterna was left open in the masonry. This is rather usual, but what is intriguing is its orientation, which is not – as it would have been easiest – perpendicular to the wall’s direction, but oblique, with a ca. 300 declination, therefore being rather a very long and tight tunnel (about 6.7 m long, 1.6 m wide). Based on material evidence, but also on architectural analogies, the gate was also dated at the end of the 3rd Century – the beginning of the 4th Century59, about the same as the generally accepted dating for Line A (!); at least two (probably partial) reconstructions of the gate are also archaeologically attested. 55 CCA 2000; Papuc, Lungu 1998, especially p. 204-7 and 202-3, figs. 1-2, 205, fig. 3 (second discovery and documentation); Lambrino 1936, with figs. 1-2 (first discovery and documentation). 56 A detail also observed by the archaeologists, see Papuc, Lungu 1998, p. 205. 57 By any rate, comparative to the eastern gate at Tropaeum Traiani, see Bogdan Cătăniciu 1979, p. 49, fig. 25. It is to notice that these long U-shaped towers are usually to be found, at least in Scythia, in gates’ configuration (typically two towers flanking the proper gate), whereas single U-shaped towers are usually smaller (see above, Line A, Tower 1) – observed here, but also at Tropaeum Traiani (eastern and western gates), Zaldapa (northeastern gate), Carsium (North Gate) – all rather large fortifications. Still, this observation needs to be checked more thoroughly. 58 According to Lambrino’s plan, 1936, fig. 2. 59 See especially the most recent dating in CCA 2000. Still, from a typological point of view, it could be dated up to the second half of the IV Century, mainly in the Constatinian period, see also Papuc, Lungu 1998. 108 The roman defensive systems of Tomis The gate is very badly preserved, especially its northern tower (Tower 4) of which the foundation was only partially found, at its rounded end. This is generally believed to be the cause of more recent damages, but there is also another possible interpretation, suggested by a picture of the Tower 3 at the date of its first archaeological uncovering60 (see Fig. 10). Its degradation is very ‘neat’, not indicating at all a natural process of decay or a ‘blind’, random damaging in the later times caused by new constructions built above; the aspect of the ruin actually indicates a systematically dismantling of the tower – and possibly, of the whole gate61. This action could have happened most likely in Antiquity62, in a time when important infrastructure works were still affordable. Line C63 was uncovered on many shorter segments distributed on two large adjoining directions: first, as an extension of Line A on ca. 365 meters, and second connecting the first and closing the city westward, spanning on ca. 330 m (at least what’s left today); it ended with the harbor’s dams, entering into the Sea. No tower has been yet properly documented on Line C or is visible today. Older references mention two possible towers on the western side’s northern half64 (Gate 4?, pointed on the plan with high probability) and a group of three rectangular towers towards the southern end of the line, in the modern Harbor’s Gate 3 area65 (Gate 5?, pointed on the plan with Lambrino 1936, fig. 1. The same thing results from the authors’ observations: “Turnul a păstrat pe întreaga sa întindere primul rând de blocuri deasupra temeliei, iar în partea sa rotundă a păstrat și cel de al doilea rând.”, Lambrino 1936, p. 914 (in free translation, only the first row of elevation blocks have been found on the sides of the tower, and two rows for the rounded side). 62 Lambrino 1936, p. 916 mentions more coins found inside the southern tower of the gate, of which the latest was from Justin (probably the first, not mentioned, 518-527), but the author also agrees that the stratigraphic context had been compromised by the further site’s use. 63 For all section referring Line C, see Cliante 2006, especially 251-2 and plan; Rădulescu 1995-1996, p. 86; Canarache 1967, p. 75, 94, 101-2; Rădulescu 1966, p. 82 and fig. 1; Canarache 1961, p. 20-1 and pl. I; Polonic mss., f. 55, 62. 64 Canarache 1961, p. 20; Polonic mss., f. 55; see also Toma 2010, p. 61 and 62, fig. 4. 65 Canarache 1967, p. 101; see also Toma 2010, p. 61 and 62, fig. 4, and PUZ 2012, where presumably one of these towers is represented. 60 61 109 Alexandra TEODOR high probability). This defensive line is considered to date in the 6th Century, as an extension of the city in the good days of Anastasius (491-518) and (up to a moment) Justinianus (527-565), although the grounds for this chronological framing aren’t conclusive at all66. The presence of two gates on defensive Line C’s western side, at approximately small distance one to another (less than 200 meters, but similar to the one between Gates 1 and 2 on defensive Line A), with implicitly two roads heading the same direction (Callatis), parallel to the shore, could seem unjustified. However, the site’s topography also has to be considered: the presumed Gate 4 is located at the upper level of the Peninsula, whereas Gate 5 is located at the Sea’s level, which is much lower (ca. 20 meters)67. Therefore, it is natural that, at least after the high cliff was terraced, a street at the harbor’s level to have also been set up, if not actually earlier, after the commercial port was established. If the street was already very important and used with intensity at the time Line C was built, as one might expect, then it is quite explainable why the ‘lower’ gate (Gate 5) was built. On the other hand, presumably the rest of the city, located on the high Peninsula, also had to be served by a connection to the extra muros area (the ‘upper’ Gate 4?), where one of the main necropolises lied along the (‘upper’) road heading to Callatis (Road B). This is exactly the situation described by Polonic, through his city plan68. The architectural documentation available for defensive line C is very scarce69; at least for the published material, archaeological evidence for Line C was only textually described and roughly marked on a general city plan. Therefore, we only know that its foundation segments are ca. 2.7-2.8 meters wide and were built from stone blocked with lime mortar mixed with sand and crushed shells. Closely to one wall fragment a large stone block was found, isolated, probably fallen; supposedly, it was part of the defensive wall, but See references and comments above, at section 4.2.1.D, and below, at 5.3. Toma 2010, p. 62, fig. 4, see level curves; Canarache 1961, p. 21 (ca. 20 meters height of the Peninsula), but also the picture at fig. 9, taken before 1904, representing the southwestern shore with the impressive ruins. 68 Polonic mss., plan, see here at Fig. 5 and also at Toma 2010, p. 62, fig. 4. 69 Basically it relies mostly on details provided by Cliante 2006, p. especially p. 250-2. 66 67 110 The roman defensive systems of Tomis one can’t know, in the light of current evidence, if it was used there for the first time or if it was a spolia. Line D70 (see also Fig. 8), marked by Tower 5 and the adjacent curtains, is on sight at least since the end of the XIXth Century, although it was never studied (?) and/or published; it may have been also mentioned by Polonic in his notes. According to RAN (“Repertoriul Arheologic Naţional”/“the National Archaeological Record of Romania”), the tower dates in the Late Roman era, but there’s actually nothing sure in this matter (no relevant bibliography or other documentation provided); the record was probably based on Vasile Canarache’s estimation. Unrelated to this, some authors actually believed this tower to have represented the southwestern end of the Early Roman defensive line crossing the Peninsula, which is highly unlikely (see also above, 4.2.1.C, Line F?). What only seems to be certain after the few more recent representations we could analyze71 is that the tower may be attached to a defensive line parallel to the shore, meant to protect the city from the very highly plausible attacks coming from the sea (see also 4.3.1); also, most likely, it was an interior tower (as Polonic also mentions for other areas on the Sea shore, see above, 4.2.1.E), a detail that would highly recommend its dating in the Early Roman period72. It was most probably reused afterwards for an indefinable time. 4.2.3. Topographical issues (see Figs. 6, 7 and 9b) The most bothering issue is the spatial relation between defensive lines A and B73. In fact, it isn’t actually known74 and this generated For all section referring Line D, see RAN, search by ‘City/Town/Village’, then ‘Constanța’, site no. 60428.31; Canarache 1967, p. 75, 86; Canarache 1961, p. 22; Polonic mss., f. 55: „În partea de sud și est[,] adică pe malul marei[,] se văde urmele unui zid de apărare de 2 m gros.” (“In the southern and eastern parts, meaning on the Sea shore, a defensive walls’ traces can be seen, having about 2 m thickness”, free translation). 71 Orthophoto ANCPI 2008, PUZ 2012; on site analysis would surely be necessary; unfortunately, it wasn’t possible at the time the study is written. 72 Lander 1984, p. 167-8, 232; Biernacka-Lubańska 1982, p. 200, 210. 73 Already mentioned before, see Teodor 2012, p. 7 (3.1.2) and fig. 5, but undetailed at the time. 74 CCA 2000; Rădulescu 1995-1996, p. 89. 70 111 Alexandra TEODOR the premise for ignoring the real possibilities for them to connect and also for promoting some most likely impossible ones. Older representations of the general defensive system included Tower 2 and its direction is approximately a prolongation of Line A’s direction; this is named here Line A, Variant 1. Of course, it was at the time, by any rate, the most obvious solution. More recently (i. e., after Gate 3/Line B was (re)discovered, in 1998), the most frequent representation is a line connecting the curtain east of Tower 1 with the curtain west of Tower 3 (named here Line A, Variant 2). But this latter option has at least two major flaws: (1) it is simply invented, since no archaeological evidence of the smallest defensive wall segment exists on its course and it is neither the prolongation of Line A, nor of Line B, therefore it’s orientation, length and starting points are absolutely random; (2) quite inexplicably, it completely ignores Tower 2, which surely exists (see above, 4.2.2). Unfortunately, this completely ungrounded hypothesis has further on become a sort of certainty in the site’s literature, never contested and also generating the most unexpected representations and interpretations75. Why was Variant 1 excluded? Well, it has never been explained, but a simple glimpse on the detailed and georeferenced plan indicates that it’s trace indifferently intersected Tower 4; this must have been really difficult to be offered an interpretation, having in consideration that the consecrated construction dating for defensive lines A and B are rather... the same. Furthermore, Line A Variant 1 not only ‘virtually’ intersected Tower 4, but – coincidence or not –, precisely on the intersection area between the two lines, Tower 4 is completely dismantled, including foundations. Truthfully, that is difficult to explain. A single plan representing both Tower 2 and Gate 3 was found so far. The PUZ 2012 authors proposed a whole new interpretation. Unfortunately, their variant is a connection between the two defensive lines that intersects at the exit point of the poterna, which was most probably misinterpreted. One also has to note that the Gate 3 to be found at Cliante 2006, p. 253, fig. 1 is misplaced by ca. 50 meters southeast, therefore generating a whole different trace of the defensive line than all previously published versions. 75 112 The roman defensive systems of Tomis 4.3. Complementary evidence 4.3.1. The stratigraphy as evidence of repeated destructions Unfortunately, from various reasons there is no stratigraphic data directly connected to the already discovered defensive elements. There were so far, in ca. 100 years of archaeological research in Tomis, only very few archaeological contexts to provide reliable stratigraphical data, all associated with residential areas, of which the most relevant is to be presented here. There were two close areas with consistent stratigraphy – Cathedral’s Park (CP) and 23 Arhiepiscopiei Street (23A), and another further away – 7 Sulmona Street (7S)76 (see again Fig. 3). Overall, more (possible) destructive fires were noticed: (1) in the second half of the VIth Century BCE (CP, level N-XI?; 23A), possibly also in the Vth Century BCE (23A, level N2); (2) at some point towards the end (?) of the 4th Century BCE (CP, level N-VIII; 23A, level N3); (3) on the level dated largely in the 1st-2nd Centuries CE (CP, level N-V)77; (4) towards the end of the 3rd Century CE (CP, level N-IV); (5?) a possible fire in the first half of the 4th Century CE (CP, level N-III); (6) towards or at the end of the same century (CP, level N-II); (7) and powerful fire at the middle of the 5th Century CE (CP, level N-I2; 23A); (8) in the first half of the 6th Century (7S, N278). 4.3.2. Related known urban infrastructure and architecture (see again Fig. 3) A key-point to many issues about the antique city’s topograohy would be a more reliable dating of the water gallery systems (watertanks)79 of Tomis. The theoretical variants for its general dating are CCA 2001; Rădulescu, Scorpan 1975, p. 11 sqq. (Cathedral’s Park); Bucovală 1998, p. 181-4 (7 Sulmona Street), where only Late Roman layers were documented. 77 See also Suceveanu 1969, p. 349. 78 This level, N2, wasn’t dated (Bucovală 1998, p. 181-2), but the next one (later, N3) was considered to represent the second half of the 6th Century – beginning of the next, see Bucovală 1998, p. 182-4. Nothing special is recorded for the oldest one, N1, and no chronological framing is directly suggested except being Late Roman. 79 Largely on this issue, Papuc 2005, especially p. 70-80 (presentation) and p. 95-8 (conclusions, dating in the Early Roman Era); for a few comments on this, see Teodor 2012, p. 3-4 (2.2.1) and figs. 1, 5-6 (3.1.1) and figs. 3-4. 76 113 Alexandra TEODOR the Nerva-Antonine (96-192) and Severan (193-235) dynasties, or later – but less likely, see also below, 5.3 –, under Diocletian and/ or Constantinus I and followers (284-361), since it was an extremly sophisticated and expensive investment. More than a dozen terminals (shafts) of this system were documented and few others are supposedly part of the same infrastructure project. One could presume that at least at its conception time this water-system had a close circuit, therefore no shaft should have been left outside the city walls, for security reasons. Few of the remotest known shafts are located at ca. 20 meters southeast of Gate 1 (P18) and presumably another (P17?)80, closely south of the large basilica located outside the hypothetical defensive Line E? – which was supposed to close the city until the 6th Century... Also, outside Line E? is another shaft (P20), located close to one of the (presumed) western gates81, Gate 5. If the last two (P17? and P20) were truly part of the gallery system – and there are many chances for them to be, since another segment of the gallery is located in the same area, near the shore82 –, then it would be a strong argument against the city walls’ western expansion scenario in the last century of the Roman city’s existence. Withal, if the water gallery system’s construction dating turns out to be in the Severan times83, then we’d have another strong argument to consider both defensive Line A and Line C much earlier than generally are now – and largely contemporary to it. Another argument for this hypothesis is the building also found above the gallery documented close to presumable Gate 5 and to shaft P20, that was dated, based on its general architecture and decoration, at the end of 2nd Centurybeginning of the 3rd Century84. Presumably, because it wasn’t mentioned whether it was an isolated shaft (i. e. reaching some natural groundwater level, which can’t be excluded) or a part of the Roman gallery system, see Rădulescu 1966, p. 53-54 and fig. 29; this ‘detail’ seemed quite certain for P18, which clearly communicated with the galleries (Papuc 2005, p. 94). 81 Polonic mss., f. 64. 82 Polonic mss., f. 64, but also documented by Tafrali 1930, p. 52-3 and figs. at pages 52-4. 83 As Papuc also suggests (2005, 97). 84 Toma 2010, p. 63-8; this building seems to be part of the south and southwestern high cliff’s systematization into terraces, as well as the so-called Mosaic Edifice. See also Canarache 1961, p. 21-22. 80 114 The roman defensive systems of Tomis 4.3.3. Numismatic iconographical data (see Fig. 11) Quite rarely is this primary source taken into account when discussing architectural issues at Tomis85, although there are rather many such representations on the local coins minted in the Roman era86. Unfortunately, the coin minting stopped with the end of Philip the Arab’s reign (249), therefore the only possible ‘evidence’ we could benefit from resumes to the Early Roman times. From our perspective, there can be pointed out at least two coin series having represented a gate (maybe the same?) with two rectangular (?) flanking towers: one is dated in the time of Geta (210-2), and another in the time of Severus Alexander (222-35)87 (see Fig. 11, top). At any rate, these iconographical representations must have had an historical ground, either a simple repair work (?) of some already existing defensive structures, either a new constructive work – more likely a whole defensive line rather than a simple gate –, one that would require imperial support. No further allegations will be made here, since our purpose was only to highlight, once more, these iconographical representations as potential arguments in the larger and complex scenario of the urban defensive system’s chronology and topographical evolution. 85 For the defensive system, this issue was brought up by Vasile Canarache (1961, p. 16 and fig. 3); still, the example presented proved not to be a gate, but more likely a triumphal arch, see also Preda 1968, p. 232, no. 57 and 236, no. 61 and here, Fig. 11, bottom. Other coins’ reverse representations were later used by Suceveanu in his own arguing for an earlier dating of the city’s precinct (here, Line A), see Suceveanu 1969, p. 348 and note54. Another use of numismatic iconographical argument (among others) related to a temple in Tomis, at Mărgineanu-Cârstoiu 2011, p. 59 and note 29. 86 City gates: Covacef Z 1980, p. 267 and pl. VIII.1; Preda 1968, p. 232, no. 50; triumphal arches: Preda 1968, p. 232, no. 57 and 236, no. 61; temples (most of the architectural representations): wildwinds.com/moushmov/tomis. A synthesis on this subject (architectural representations on coins), for Dobruja, would be extremely useful. 87 For the Geta coin, the published image has a very bad quality, see Covacef Z 1980, p. 267 and pl. VIII.1, therefore no comparison was possible; for Severus Alexander coin, see Preda 1968, p. 232, no. 50, reproduced here at Fig. 11, top. 115 Alexandra TEODOR 5. Discussions and Hypotheses (see again Fig. 3, 6-10) 5.1. Defensive Line A. As seen above, in section 4.1, although the Early Roman necropolis’ limits are rather unclear – especially because not few are the cases when the graves are dated on a large span, say 1-2nd or 2nd-3rd Centuries (a ‘detail’ which at some point can be crucial), but also because not all funerary evidence are thoroughly mapped when published88 –, still apparently no contemporary funerary discovery seems to cross (towards the city) the southern and western limits marked in Fig. 4 for the 2nd-3rd Centuries necropolis (NL2 and NL3)89. This allows us to consider that defensive Line A theoretically could have been contemporary to the 2nd-3rd Centuries necropolis. Besides the ‘consecrated hypothesis’ regarding defensive Line A’s construction date (see above, 3), there is also another – largely disregarded –, that dates it much earlier than usually considered, i. e. towards the beginning of the 2nd Century90. It relied on the funerary evidence at that time91, on the numismatic representations (see also above, 4.3, notes) and generally on the grounds of the city’s great development in the Early Roman times, seen through – among others – the rich epigraphic evidence and its status as Metropolis of the Western Black Sea Greek cities’ community. The whole hypothesis was later easily dismissed after the discovery of a funerary area dated most probably in the first half of the 2nd Century, located inside Line A92, It is only a general observation made on the brief process of mapping the most relevant funerary evidence; an inventory of such cases wasn’t intended, though. 89 The same observation at Suceveanu 1969, p. 348. 90 Suceveanu 1969, p. 349. It wasn’t rejected by Georgescu 1998, p. 89 and notes 62-5, still the ‘consecrated’ one was preferred, see p. 78-9. It was admitted by Toma 2010, p. 72 passim, though suggesting a later date, in second half of the 2nd Century. 91 Barbu 1961 and Bucovală 1966, see Suceveanu 1969, p. 348, n. 53. Still, the 2nd Century cremation grave found on Marc Aureliu Street, published by Zirra, Alexandrescu 1957, p. 89 was ignored by all quoted sources. 92 In the sector south of Gate 3, where a large (episcopal?) basilica was found, partially overlapping an amphitheater, see Rădulescu 1991, p. 28 88 116 The roman defensive systems of Tomis and the discovery of some coins dated in the last quarter of the 3rd Century in the emplecton of Line A’s curtain walls – ignoring, as previously mentioned, the possibility that they could simply indicate a reconstruction. However, besides the funerary evidence mentioned above, that so far allows an earlier dating than the consecrated one, the details regarding the constructions found in the presumed Gate 5 area are key points in this issue. The marble finished building on the coast line, near the named gate, was dated towards the end of the 2nd Century – first part of the 3rd Century; together with the so-called Mosaic Edifice ensemble, these constructions point out the high cliffs’ unitary systematization (at least) along the whole southwestern shore. Both buildings were constructed above the sophisticated water gallery system having terminals (shafts) spread all over the surface closed by defensive Line A. Last, but not least, Line A gates’ (1 and 2) typology and construction system speak broadly of an earlier defensive work than the ones an imperial army would have built at the end of the 3rd Century. We consider all these are strong arguments for the earlier dating of Line A than usually considered, except not at the beginning of the 2nd Century, but rather towards the end of the 2nd Century – first part of the 3rd Century. Its construction may have been necessary as an effect of the general development from the Severan dynasty’s period, that followed the Costoboces’ devastating invasion from 170 (presumably to also have affected our city and implicitly the city wall, see 4.3.1, no. 3); therefore, it is likely that a larger defensive line was needed at these times. 5.2. Defensive Line B. We’ve seen that Line B strongly indicates a different topographical approach (including a different orientation) of the defensive walls compared to Line A, and also a different military architecture. The fact that Line B’ direction is practicaly overlapped by a modern street is also worth of giving a thought, afterall Ferdinand Bulevard’s topography is, for most of its northeastern course (and implicitly the southwestern), nothing but (for Suceveanu’s hypothesis rejection and graves’ dating), 29-34 and fig. 4-6 (basilica), 30, 35 (amphitheater). See a possibly more refined graves’ dating (end of first Century, first half of the 2nd Century ) at Georgescu 1998, p. 71 and also p. 60 for the spatial references used there – if the author refers to the same funerary area, which is uncertain. 117 Alexandra TEODOR a sequel of the defensive Line A’ presence93. On the other hand, there is no doubt that Line B was built sometimes at the end of 3rd-first half of 4th Century, therefore (according to our chronological perspective) almost a century after Line A was probably built. Line B’s Gate 3 seems to have functioned for some long time, but ultimately it has probably been dismantled, most likely still during Antiquity (see above, section 4.2.2). Supposedly in the 6th Century, Tower 1 was (re) built on Line A, a fact that actually proves their functioning in this period – also attested by the two Christian basilicas behind Line C. This leaves us with one big question and also few others correlated, at this point all rather difficult to answer: Q1.Was Line B part of a different defensive system than Line A? If the answer is yes, Q1.A. What generated the city’s defensive line reduction? Q1.B. What was Line B’ course and where did it close? Q1.C. What happened to Line A when Line B was built and while it functioned? Q1.D. Why later was Line B dismantled (if it truly was) and what generated the city’s (new) expansion? If the answer is no, Q1.E How could Tower 2’s presence be explained in relation to Gate 3? Could they have actually functioned together from a military (defensive) point of view? Q1.F Subsequently, from an architectural and topographical point of view, how was their joint resolved? No one can offer a final answer to any of these questions in the current state of knowledge, but only hypotheses meant to be defended or dismissed, later on. Gate 3 is a reflection of the late 3 rd-4 th Century defensive architecture; undoubtedly, it was an imperial construction, meant to protect the new provincial Capital and to highly represent imperial power, probably being the biggest gate of its type in Scythia. Why Ferdinand Bd. and the parallel street north to it (with houses between them) overlap what probably used to be the ditches area in front of the city walls or even a street/road parallel to it, see Polonic’s plan at Fig. 5. For the antique heritage in the modern city’s street network, see also below, 5.5. 93 118 The roman defensive systems of Tomis the imperial authorities would have chosen a different defensive line than previous – if they did (Q1.A)? The only explanation we could offer at this point is that it could have been preferred a shorter, straight line (Q1.B: ca. 700 meters on Line B’s direction, most plausible to be found on today’s Dragoş Vodă Street, see also below, 5.5), but better defended, compared to a longer, multiple braked one (ca. 1100 meters on Line C-Line A’s traces), and weaker in front of the potential enemies. Also, from a technical point of view, it could have probably been preferred a new construction (a fundamentis), rather than a complicated adaptation of new military architecture to older one. To sum up, it could have been a pragmatic option. If this actually happened, then the space between Line B and Lines A and C (?) could have remained residential, since no funeral evidence was found for the Late Roman Era (Q1.C); probably Lines A and C weren’t even destroyed. Further on, why the 6th Century authorities would have given up on Line B (since Gate 3 was dismantled) and have returned to Lines A and C (since Tower 1 was now (re)built) (Q1.D)? Besides a possible destruction of the city at the middle of the 5th Century (see below), other causes should be identified in order to justify the presumed enlarged area. As for the possible construction and reconstruction occasions (again, reffering to Q1.A and D), we’ve seen there were archaeologically attested no less than six, maybe eight, serious distructions during the whole Antiquity, at least in some parts of the city (see 4.3.1). Although the few points where these destructions are attested94 are not necessarily relevant for all city, they are – for now – the only physical evidence to count on. Three (or maybe four) destructions occured most probably after the Line A was built and no less than two (or maybe three) occured most probably after Line B was built (therefore, one clear destruction between their presumed construction time, in our view). It is therefore a possibility that the premises for Line B were due to the fire dated towards the end of 3rd Century (idem, no. 4), which could have altered Line A. Most likely at the time of the destruction attested in the 6th Century (idem, no. 8) See also gzt.ro-3 2013 for indications regarding a possibly general fire destruction in the Late Roman Era. 94 119 Alexandra TEODOR Line A and C were functional, therefore this leaves us with the most plausible Line B’s destruction caused by the violent attacks in the middle of the 5th Century (idem, no. 7). Given its neat appearance, it was probably dismantled during the massive construction works associated with Anastasius and Justinianus’ reigns, when apparently efforts were concentrated on reconstructing or simply repairing the defensive Lines A and C and also of the nearby large basilicas. Regarding the question Q1.E, other authors have already expressed their opinions tending towards a negative answer95; we couldn’t find any arguments in the opposed direction. Question Q1.F could only be answered based on more precise topographical data and, ideally, by some more consistent archaeological evidence. 5.3. Defensive Line C. Apparently, defensive Line E? which was considered to have closed Line A westwards has no archaeological ground. Furthermore, Line E?’s hypothetical direction is largely incompatible, from a topographic point of view, with the water gallery network it would probably cross over at the lower levels, also leaving outside P18 (highly probable a gallery systems’ shaft). There is also another basic topographical incompatibility to be noted: the ca. 20 meters level difference between the high shore and the sea level are supposed to be cut by the city wall on an angle of ca. 500, which is highly unlikely at least from a constructive point of view. It is important to note that if this defensive line never existed, then we’ll have to answer to the question: Q2. Where did the city walls close on the western side before the 6th Century? Well, no other defensive walls than Line C’s segments are yet uncovered in this direction... On the other hand, the only few available architectural details regarding Line C (the apparent high Papuc, Lungu 1998, p. 207, though the situation was most probably ambiguously understood; the author quotes the same source (Canarache 1961, p. 20) referring at – presumably – two different towers: the one east to Tower 1, that is Tower 2, as argument to consider Gate 3 contemporary to Line A; and another (?) tower, at ca. 100 meters northwest (actually, west) to Gate 3, considered incompatible with it (names according to our conventions); but Canarache 1961, p. 20 only mentions one tower alone in this area, and that is our Tower 2. 95 120 The roman defensive systems of Tomis quality mortar, for example) also point out towards an earlier dating than 6th Century; furthermore, no funerary evidence known so far crossed this defensive line later than 1st-2nd Centuries. Therefore, one can count already enough arguments to at least hypothetically consider Line C contemporary and, basically, one and the same with Line A. 5.4. Defensive line towards the Sea. It is highly probable that the wall segments named here D and H were part of the same defensive line, i. e. the one that surrounded the (upper) city, towards the Sea. In this area the defensive line most likely didn’t change much. We can’t precisely date them in the current light of knowledge, but one can assume, being known their fortification elements’ typology (interior rectangular towers) and construction system (opus quadratum, dry assemblage), that they are likely to having been built in the Early Roman times (possibly contemporary to Line A and probably C), eventually at some parts (the southern ones) superimposing older fortifications (maybe Greek). On the other hand, one could easily observe that all the up mentioned segments (D and H) are located on the high level of the Peninsula, therefore leaving the warehouses and the rich buildings above them outside the presumed city wall, exposed to possible attacks from the Sea96. If the early dating hypotheses (Severan age) of Lines A and C, and also of the shore buildings turn to be right, a question that needs to be answered is: Q3. What was the spatial relation between them, or how did they function together, without altering their main purposes?97 Still, these buildings on the shore probably weren’t completely exposed to the Sea. Besides the antique traces of a dam observed by Polonic as a prolongation of Line C towards the Sea, there may 96 To note that Cathedral’s Park sector is almost 1 km away from the closest known gate of the city (Gate 3 or ‘The Main Gate’, see Fig. 3), but only 100-200 meters away from the Sea; the Sulmona Street sector is ca. 400 meters away from the same Gate 3, but also much closer to the sea (see Fig. 3). This means that either most of the city was destroyed, either it was attacked (also) by the Sea – if the fire was the result of an attack. See also 4.3.1. 97 See also Rădulescu 1995-1996, p. 91-2. 121 Alexandra TEODOR be relevant the (only) underwater investigations carried so far and published some time ago by Constantin Scarlat98. They both point out that the southern shore was most probably screened by a dam closing an artificial harbor. Judging by the spatial relation between the dam seen by Polonic and Line C, it is likely that they were part of the same project. 5.5. The gates and the presumable street network (see Fig. 12) Theoretically, as already mentioned, necropolises’ limit lines towards the city would mark the remotest possible position of the contemporary defensive line and implicitly the gates’ remotest possible location. As mentioned before (for all, see above, 4.1), there were identified three such possibly relevant limits: NL1, delimiting the 1-2nd Centuries funerary area – eventually also the Greek one –, and NL2-NL3, delimiting the 2 nd-3 rd Centuries funerary area northwards and westwards – most probably also the Late Roman funerary limits. Behind those limits (i. e. towards the city), along the hypothetical road lines were marked the remotest theoretical gate locations: G1, for the first part of the Early Roman Era (also Greek?), along Road A; for the second part of the Early Roman era, G2 along Road A, G3 along Road D, and G4 along Road C. Quite closely to G3 we find the real Gate 1 (ca. 60 meters), and very close to G4 the real Gate 2; still, it’s not so much the small distance that’s relevant, but the presence of the real gates on these road lines that, according to funeral evidence, date back to the 2nd-3rd Centuries or maybe earlier. True, it’s difficult to say which was first – the gate or the road connecting it to the territory? So we may assume that the roads were already used when the city fortified area grew, and city gates were placed along them; therefore, the actual (largely) terminus post quem is the above mentioned time interval, and not the coins found in the wall of defensive Line A. It’s not completely excluded for a third gate to correspond to point G2 on Line A, at its intersection to Road A, since the last was probably the oldest and most important road connecting the city to its territory. Scarlat 1975, p. 257-72 and fig. 8; see also comments at Teodor 2012, p. 3-4 (2.2.1) and fig. 1, also 5-6 (3.1.1) and figs. 3-4. 98 122 The roman defensive systems of Tomis Defensive Line B’s Gate 3 is located at about 70 meters eastwards from the theoretical limit (G2) marked on Road A, and the gate’s axis is at about 50-60 meters northeastwards to the road’s hypothetical axis, therefore it is possible for them to actually have overlapped in reality, meaning the road’s axis could have passed through Gate 3; still, that is not necessary, since Gate 3’s street/road could have been a simple variant to Road A. In this matter, we’ve found some indirect clues in the following hypothesis: that in the northern half of the antique city many road/street variants (parallels) existed back then quite approximately as they also exist today, in the modern city. This hypothesis resulted from comparing relations between antique related elements (both theoretical and real) with the actual street network. An interesting relationship was observed between the already known gates (gates 1-5) and the theoretical ones without correspondent (G1, G2). It is to be expected that when the city developed, say from the defensive Line G? to the defensive Line A, the old theoretical city gate (G1) was disposed; still, its place must have been quite an important spot, firstly because it was located in a very tight area of the Peninsula (presumably somewhere on the northwestern line of the Ovidiu Square). Therefore, we can assume it could have been the starting point to the street network tracing of the city’s new area, connecting it with the new gates. Two, but maybe three or even four new gates opened through the new city walls (Line A and presumably Line C); therefore, most probably the old gate’s spot (G1) was connected to the new gates through streets, as they were before connected through roads. Let’s see how these streets could have lined up. We’ve connected theoretical point G1 to the known real gates (1, 2 and 3) and to the theoretical gate point, G2, since it has no real correspondent known so far; there resulted four hypothetical streets (from west to east, named by the correspondent road): Street 1C (G1Gate 1); Street 1D (G1-Gate 2); Street 1A (G1-G2); Street 1E (G1-Gate 3). We have yet no archaeological evidence to sustain these street paths, but there’s something else that may count. If we take a look at the 123 Alexandra TEODOR up mentioned street paths overlapped on today’s street network, we may notice the following99: - Street 1C is closely parallel to Traian Street (its southeastern end), at about 10-12 meters north; they run parallel for about 280100 meters starting from Ovidiu Square’s northwestern line, after which our hypothetical street runs free for about 460 meters; - Street 1D is practically coincidental to Tomis Bd. for about 370 meters starting from Ovidiu Square’s northwestern line and is doubled by other urban features for about another 100 meters, then runs free for another 200 meters; - Street 1A is possible to approximate to Ecaterina Varga Street’s course, although the latter follows a slightly broken line; it doesn’t start from Ovidiu Square, but about 200 meters north from it. Still, our hypothetical street runs along today’s street for all of its course, measuring up to ca. 440 meters, until it meets Ferdinand Bd. (i. e. approximately the line of the former defensive Line A); - Finally, Street 1E has no correspondent in the present street network; still, if there’s something to learn from the previous three correspondences, then we could assume that G1 may have been connected to Gate 3 through a hypothetical street following roughly today’s Mircea cel Bătrân Street. Apparently, this would mean that in the northern half of the antique city, what is now a (major) street may as well have been in the Roman times. Therefore, if the defensive Line B is now roughly overlapped by Dragoş Vodă Street101, it’s course may as well have been a street in the 6th Century – an argument that enforces the assumption of Gate 3’s dismantling in Antiquity102. Basically similar (empirical) observations related to the antique street network, at Georgescu 1998, p. 92-3. 100 All distances are measured in straight lines. 101 See also Toma 2010, p. 58 and 62, figs. 2 and 3 for plans representing old medieval city’s street network, containing also this street. 102 A partially similar case is known at Histria, where during the Late Roman period a part of the former Hellenistic southern and northern defensive lines (initially reused) were transformed into streets after the city walls’ extension in both directions, see Domăneanţu, Sion 1982, p. 379-84 (phases A and B). 99 124 The roman defensive systems of Tomis Related to Gates’ 4 and 5 connections to the intra muros street network, the site’s topography doesn’t leave many options. Starting from Gate 4 there probably was a street having similar direction to Traian Street (its middle-eastern course), towards the center and probably intersecting Street 1C; we’ve called it here Street 1B. From Gate 5, most likely the harbor’s street had a similar trace to our days’ Portului Street; we’ve called it here Street 1B’103. The established relationship between the hypothetical antique street network and the modern one for the area analyzed is, if not surprising, at least encouraging to the idea that these streets’ baseline topography didn’t change much from Antiquity. If the assumption is true, then we could consider having also a theoretical confirmation for the Early Roman gate to be found somewhere close to the theoretical one, G1 – one of the starting points in our investigation regarding the street network. 5.6. The Roman agora? (see again Fig. 12) Continuing last point’s hypothesis, if the former (theoretical) gate location G1 was at the time when defensive Line A was built (and ever since) the intersection of three, maybe four streets, it must have been quite an important site of the enlarged city (as it is today, see the plan). Its formerly assumed importance and potential role as an agora104, being given the special location in the tightest area of the Peninsula and the near presence of the so-called Mosaic Edifice (in fact, most probably a civil basilica), later also possibly a Christian basilica on the opposite side of the square105, is now reinforced by the arguments regarding the street network concentrating in this area (see above, 5.5). 103 Obviously, this implies a correspondent Road 1B’, parallel and south to Road 1B, both heading to Callatis; it was not represented on Figure 4 because it wasn’t determined or supported by funerary evidence – at least, not in the city walls’ proximity. 104 Toma 2010, p. 69-70; Canarache 1966, p. 3-4. It doesn’t mean this was also the Greek agora. If the defensive line in that period really was around Line G?, it is less likely because their proximity, therefore it should be searched somewhere in the southern half of the Peninsula. 105 Lungu 2000, p. 69-70; see here on plan at Fig. 3, point 10. 125 Alexandra TEODOR Lastly, it is important to note here two other details that may be connected to this place – as mentioned, possibly the city’s agora. First, is that recently a solar clock was discovered in the Ovidiu Square area106, also underlining the spot’s high public importance in the city’s life (assuming its discovery location was close to the place where it was originally installed). Second, there is a very interesting numismatic evidence – also rather ignored by authors concerned by urban Tomis – representing a monument that only now can be given a potential location: a coin series dated in the time of Maximinus Thrax (235-8), representing most likely a triumphal arch with three openings. Most likely, its impact would have been best achieved by locating it at the agora’s main entrance(s) coming from the city’s gates, somewhere at the intersection of the upmentioned streets, in the present Ovidiu Square area. 6. Further Research and Good Practice Instead of systematicaly pointing out research directions for each of the raised issues here, which are somehow implicit (most of them require archaeological verification), I prefer listing only the large, general directions for this site. 6.1. A precise topographic digital documentation of all unearthed ruins of the antique city and all the ones that will be uncovered in the future must be made107. The lack of a precise topographic support may be one of the main causes for all faulty associations between the known defensive elements. All the above mentioned hypotheses should also be checked108 after the precise topographic data will be collected. gzt.ro-3 2013. Such an archaeological documentation should exist, indifferent to the site’s status as a systematically or rescue investigated archaeological site. Moreover, this documentations’ existence publicly (online, for free use) would be very useful for both research and protecting the modern city’s archaeological heritage. 108 Most of the key elements discussed here – Line A and B walls’ location and orientation – were mapped as precise as possible given the available data – published plans for gates 1 and 3 and Tower 1, overlapped on georeferenced good quality orthophoto. Still, undoubtedly our topographic data must be 106 107 126 The roman defensive systems of Tomis 6.2. A thorough digital layered integration of all known (both published and yet unpublished) topographical data regarding the antique (eventualy medieval and modern) city must be done in order to enable one to customly analyze different urban aspects of the city, starting from the defensive systems, infrastructure (terraces, water systems, drainage etc.), ending with the smallest residential complexes. All these elements are intimely connected one to another, therefore the whole relational system must be understood in order to get ‘the big picture’ related to the city’s evolution. Where specific topographic data (drawn) isn’t available, all textual references should be mapped as precise as possible. 6.3. Having the above points satisfyingly checked, further interpretations and deductions have to be made in order to better understand all the integrated data, ideally in a digital multilayered environment109. A thorough hypotheses’ evidence must be kept, in order to be easily revisited, as in a programmers pipeline: ‘if-thenelse’ (briefly illustrated above, see 5.2), as complex at it may be. 6.4. Whenever possible, if adequate, surface investigations (largely, geophisycal in this case) should be made in order to systematically filling the puzzle. Also, any archaeological rescue mission in the Peninsula should use the (presumable) data gathered in order to conceive a project-based research task meant to squize all possible information one could list for the offered spot. The antique city actually can reveal itself, with one condition: really wanting to see it and using all available tools and tehniques for this purpose. Only after these points will be more or less covered, we’ll be entitled to say that Tomis’ remains are properly documented and investigated for the XXIst Century. Furthermore, any responsible digital reconstruction of the antique city or its elements should be connected to the above listed points and its details ‘proportional’ to the acquired knowledge. Last, but no least, no restoration project confronted to another data set, ideally georeferenced digital topographic measurements, in order for its accuracy to be improved. 109 The software market is very generous, therefore the main condition in any such project should be the data set’s interoperability (meaning especially georeferenced and vector data), indifferent to one’s working environment. For our project’s details, see above, section 2 and notes. 127 Alexandra TEODOR should begin before the detailed archaeological, topographical, architectural and related documentation is done, as it had unfortunately happend for defensive Line A’s uncovered elements. 7. Conclusions Although some of them uncovered for a long time now – and others more recently –, the defensive elements of Tomis don’t seem to have been precisely documented and, therefore, were most probably often misinterpreted. This insight was meant to reopen some ignored hypotheses and to open new ones regarding the defensive elements’ topography and roughly their chronology, and to confront them to the consecrated ones. Both the study’s approach and its presentation followed the same methodological steps, seeking and underlining connections between funerary evidence – possible roads – known/ possible defensive lines and gates – possible street network. One of the main conclusions is that the so-called Late Roman defensive line’s construction date is most likely earlier than generaly considered (here, Line A). As previously noticed already, it could date with high probability in the Severan era, as well as the terracing along the southwestern shore, best documented at the so-called Mosaic Edifice, but also other infrastructure works, such as the water gallery system extended throughout whole its intra muros area. Compared to the recent and similar chronological allegations mentioned, here were assembled some suplimentary arguments for its earlier dating (for instance, the gates’s typology and the construction details published by Bănică, basically ignored previously, as well as their correspondence to the segments on the Sea shore). Few ‘details’ left undiscussed were also pointed out. First, it is the so-called 6th Century city wall extension (here, Line C), which most likely is one and the same with the here supposed Severan defensive line (Line A). Second, the questionable topographical relation of the latter with the so-called Main Gate (here, Gate 3), an undisputable Tetrarchic or Constantinian military construction. Third, the highly plausible Gate 3’s dismantling during Antiquity and, therefore, the hypothetical complications deriving from this. 128 The roman defensive systems of Tomis Another conclusion is that the city most probaby did have – at least for some while – city walls towards the Sea, on the high level of the Peninsula, therefore being protected by a closed defensive line configuration; by its typology and construction system it is to be dated the latest in the Early Roman period. What has to be established is for how long did the shore defensive walls function and what was their spatial and chronological relation to the buildings on the shore. Also, it seems quite clear, as previosuly stated, that the Roman city had an artificial harbor; we might add that it was probably part of the same project as the city walls (Lines A and C) and the terracing system along the shore. If we sum up, the inventory of elements which could be assigned to the Severan period (the artificial harbor, the most extended city walls, the water gallery system, the high shore’s terracing and the corresponding public buildings and warehouses) leads us to the conclusion that this period was, at least from a constructive point of view, the peak of this city’s development. Therefore, by its configuration and infrastructure, Tomis is an authentic Early Roman city, later on more or less modified, mostly at the built stock level and mainly by the emergence of the Christian official architectural program. One last conclusion regards the antique main street network; though its possible configuration has been previously empirically inferred, here its (hypothetical) confirmation was an indirect result to the methodological deductions followed. It seems that, despite the already noticed fact that in the southern Peninsula the street network most probably changed in the modern times110, in the northern area of the antique city the main street network may have actually been extremely well conserved up to our days; this observation was probably the main input of the methodological approach. Teodor 2012, p. 7-8 (3.1.3) and figs. 6-7; Coja 1982, fig. 221. A different opinion at Toma 2010, p. 72 and 62, fig. 4. 110 129 Figure 1. (Left) Presentation panel next to the northern gate (Gate 2, here), in September 2013, saying: “Tomis city’s fortifications, with two gates [sic!] and more defense towers. The western gate [Gate 1, here, but not the one pictured] is flanked by two towers and still preserves the stepping area and its framework. Through it, the connection with southern littoral cities was made. / The northern gate [the one pictured], 2.5 meters wide. It connected Tomis to the regions’ northern and inside territories.” Actually, the visitor would really need to have good orientation skills in order to realize which of the gates one enters. (Right) Entrance to the northern gate, rebuilt similarly to the western one – for which more archaeological data was available (see text, 4.2.2); its height is neither slightly completing the ruins, neither offering the realistic reconstructed volume, but it’s just perfect for children’s play. Alexandra TEODOR 130 Figure 2. Virtual reconstructions (VR) of the city’s defensive system: (VR1, left) cerva.ro/tomis/; (VR2, right) cjc.ro (screenshots). Both VR ignore Tower 2 (see the text, 4.2.2). In VR1-top, the harbor has no protection (dams) and the high shore’s (sic!) terraces are only represented for the so-called Mosaic Edifice, leaving the rest of the shore area green; besides the fortification lines that generally respect the known defensive elements, the urbanistical restitution (VR1-middle and bottom) is pure imagination, without streets and each building having its own orientation (!). VR2 is more ‘fair’ (not representing the unknown), but is faulty at least for Line C’s representation (top: no apparent angle between northern and western sides, see also Fig. 3). The roman defensive systems of Tomis 131 Figure 3. Tomis. Defensive elements (archaeologically attested and hypothetical) and places mentioned in the text. For references see the text, section 4. Alexandra TEODOR 132 Figure 4. Funerary evidence (4th Century BCE – 7th Century CE), late 19th Century visible (funerary?) mounds (apud Polonic mss., see Fig. 5); possible road axes indicated by the funerary evidence’ distribution; and possible gate locations on the route roads considering necropolises’ known limits. For references see the text, section 4.1. The roman defensive systems of Tomis 133 Figure 5. Pamfil Polonic’s plan (Polonic mss.): (a, left) georeferenced, overlapped with the present topographical plan, see legend and scale attached; (b, top-right) the original, no precise scale, to be compared to a. The georeferencing process implied some specific ‘deformations’ of the original plan, best visible at the plan’s borders (straight for the original, ‘deformed’ for the georeferenced one). Alexandra TEODOR 134 Figure 6. Defensive Line A on orthophoto (ANCPI 2005). For references see the text, 4.2.2. The roman defensive systems of Tomis 135 Figure 7. Defensive Line B on orthophoto (ANCPI 2005) and possible connections to Line A. For references see the text, 4.2.2. Alexandra TEODOR 136 Figure 8. Defensive Line D on orthophoto (ANCPI 2008) (left) and detail (right). For references see the text, 4.2.2. The roman defensive systems of Tomis 137 Figure 9a. Architectural details overlapped on orthophoto (ANCPI 2008), Line A. Detailed plans for Gate 1 (left, Bănică 1991, 64, fig. 9) and Tower 1 (right, Pârvan 1915, 440, fig. 4). For further references see the text, 4.2.2. Alexandra TEODOR 138 Figure 9b. Architectural details overlapped on orthophoto (ANCPI 2008), Line B. Plans for Gate 3: yellow, Lambrino 1936, p. 913, fig. 2; magenta, Papuc, Lungu 1998, p. 203, fig. 2. For further references see the text, 4.2.2. The roman defensive systems of Tomis 139 140 (RIGHT) Figure 11. Architecture represented on coins minted in Tomis: top, city gate (apud Preda 1968, p. 234, no. 50); bottom: triumphal arch – left, unidentified (apud Canarache 1961, p. 16, fig. 3); right, from Maximinus Thrax (235-8) (apud Preda 1968, p. 234, no. 57a). See also the text, 4.3.3 and 5.6. (LEFT) Figure 10. The southwestern tower of Gate 3 (Tower 3) at the first discovery, in the 1930’s (Lambrino 1936, fig. 1); to be noticed the neat aspect of the ruin, most probably the result of the walls’ systematical dismantling. See also the text, 4.2.2. Alexandra TEODOR Figure 12. Presumed Roman street network’s axes, as resulted from the study, and their spatial relation with the present street network; see the text, 5.5. The roman defensive systems of Tomis 141 Alexandra TEODOR Bibliography Vasile Barbu, Considération chronologiques basées sur les données fournies par les inventaires funéraires des nécropoles tomitaines, Studii Clasice, 3, 1961, p. 203-25. Băjenaru, Constantin Băjenaru, Cătălin Dobrinescu, Săpături de Dobrinescu salvare în necropola romană a Tomisului, Pontica, 41, 2008, 2008 p. 189-99. Bănică 1991 Teodor Bănică, Tomis: Poarta nr. 1 a incintei fortificate, Revista Monumentelor Istorice, 2, 1991, p. 62-65. BiernackaMałgorzata Biernacka-Lubańska, The Roman and EarlyLubańska 1982 Byzantine fortifications of Lower Moesia and Northern Thrace, Wrocław, Polish Academy of Sciences, 1982. Bogdan Ioana Bogdan Cătăniciu, Incinta. Structura, datare, istorie, Cătăniciu 1979 in Ion Barnea et al. (ed), Tropaeum Traiani I. Cetatea, Bucureşti, Editura Academiei R. S. R., 1979, p. 47-63. Bucovală 1967 Mihai Bucovală, Necropole elenistice la Tomis, Constanţa, Muzeul Regional de Arheologie Dobrogea, 1967. Bucovală 1968 Mihai Bucovală, Noi morminte de epocă romană timpurie la Tomis, Pontica, 1, 1968, p. 269-306. Bucovală 1975 Mihai Bucovală, Un alt mormînt din epoca elenistică la Tomis, Pontica, 8, 1975, p. 375-88. Bucovală 1995 Mihai Bucovală, Un alt mormânt elenistic descoperit la Tomis, Pontica, 28-29, 1995-1996, p. 73-82. Bucovală 1998 Mihai Bucovală, Raport preliminar privind cercetările arheologice cu caracter de salvare din Constanţa, strada Sulmonei nr. 7, Pontica, 31, 1998, p. 171-200. Bucovală, Pașca Mihai Bucovală, Cecilia Paşca, Descoperiri recente în 1988-1989 necropolele de epocă romană şi romano-bizantină la Tomis, Pontica, 21-22, 1988-1989, p. 123-61. Bucovală, Pașca Mihai Bucovală, Cecilia Paşca, Descoperiri recente în 1991 necropola romană de sud-vest a Tomisului, Pontica, 24, 1991, p. 185-236. Bucovală, Mihai Bucovală, Cecilia Paşca, Cercetări în necropola romană Paşca 1992 de vest a Tomisului (1992), Pontica, 25, 1992, p. 241-72. Buzoianu, Livia Buzoianu, Maria Bărbulescu, Tomis. Comentariu Bărbulescu istoric și arheologic / Historical and Archaeological 2012 Commentary, Muzeul de Istorie Națională și Arheologie Constanța, 2012. Barbu 1961 142 The roman defensive systems of Tomis Canarache 1961 Vasile Canarache, Tomis, Bucureşti, Meridiane, 1961. Canarache 1966 Vasile Canarache, Edificiul cu mozaic din Tomis, Bucureşti, 1966. Canarache et Vasile Canarache, Adrian Rădulescu, Andrei Aricescu, and al. 1963 Vasile Barbu, Tezaurul de sculpturi de la Tomis, Bucureşti, 1963. CheluţăNicolae Cheluţă-Georgescu, Contribuţii la topografia Georgescu 1977 Tomisului în sec. VI e. n.,Pontica, 10, 1977, p. 253-60. CheraConstantin Chera-Mărgineanu, Noi morminte din necropolele Tomisului, Pontica, 12, 1979, p. 247-250. Mărgineanu 1979 CheraConstantin Chera-Mărgineanu, Virgil Lungu, Contribuţii Mărgineanu, la cunoaşterea unor necropole creştine ale Tomisului (II), Lungu 1983 Pontica, 16, 1983, p. 217-30. Cliante 2006 Laurenţiu Cliante, Un nou segment al incintei tomitane târzii, Pontica, 39, 2006, p. 249-58. Coja 1982 Maria Coja, Les fortifications grecques dans les colonies de la côte ouest du Pont Euxin, in La fortification dans l’histoire du monde grec, P. Leriche, H. Tréziny (eds.), CNRS Colloque International, vol. 614, 1982, p. 95-103. Constantin et Robert Constantin, Laurențiu Radu, Mihai Ionescu, and al. 2007 Nicolae Alexandru, Mangalia. Cercetări arheologice de salvare,Peuce, S. N., 5, 2007, p. 241-96. Covacef P 1998 Petre Covacef, Despre un tumulus la Constanţa, Pontica, 31, 1998, p. 261-4. Covacef Z 1980 Zaharia Covacef, Raport preliminar asupra cercetărilor arheologice din sectorul V al cetăţii Capidava (campaniile din anii 1975, 1976, 1978 şi 1979),Pontica, 13, 1980, p. 254-74. Covacef Z 2011 Zaharia Covacef, Sculptura antică din expoziția de bază a Muzeului de Istorie Națională și Arheologie Constanța, Cluj-Napoca, Editura Mega, 2011. Domăneanţu, Catrinel Domăneanţu, Anișoara Sion, Incinta romană tîrzie Sion 1982 de la Histria. Încercare de cronologie, Studii și Cercetări de Istorie Veche și Arheologie, 33, 4, 1982, p. 377-96. Georgescu 1998 Nicolae [Cheluţă-]Georgescu, Urbanistica romanobizantină în aria vest pontică (sec. IV-VII e.n.), Universitatea „Al. I. Cuza”, Facultatea de Istorie, unpublished PhD thesis, 1998. 143 Alexandra TEODOR Lambrino 1936 Scarlat Lambrino, Turnul și poarta cea nouă ale cetății Tomis, Arhiva pentru ştiinţă şi reformă socială, II, 1936, p. 912-7. Lander 1984 James Lander, Roman Stone Fortification: Variation and Change from the First Century A.D. to the Fourth. University of California, Los Angeles, 1984 (published, BAR International Series 206, 1984). Lungu 2000 Virgil Lungu, Creştinismul în Scythia-Minor în contextul vest-pontic, Sibiu-Constanţa, 2000. MărgineanuMonica Mărgineanu-Cârstoiu, Némésis et la Coudée. Un Cârstoiu 2011 édicule votif de Tomis, Caiete ARA, 2, 2011, p. 53-68. Michel 1862 J. Michel, Les traveaux de défense des Romains dans la Dobroudja, Paris, Société Antiquaires de France, 1862. Papuc 2005 Gheorghe Papuc, Tomis I. Aprovizionarea cu apă a cetăţii Tomis în epoca romană târzie, Constanţa, Ex Ponto, 2005. Papuc, Lungu Gheorghe Papuc, Liviu Lungu, Poarta mare a cetăţii Tomis, 1998 Pontica, 31, 1998, p. 201-8. Pârvan 1915 Vasile Pârvan, Zidul cetății Tomi, Analele Academiei Române: Memoriile Secțiunii Istorice, 2, 37, 1915, p. 416-436. Polonic mss. Pamfil Polonic, Ruinele oraşului Tomis (Constanţa), Academia Română, Fond Pamfil Polonic, Mss., mapa 3, plic I, caiet 9, p. 51-77 and plan. Popovici Dorin-Mircea Popovici, Valentina Voinea, Ciprian Ilie, Mihai et al. 2010 Polceanu, Victor Rizea, and Vera Hramco, Reconstituirea virtuală a edificiului roman cu mozaic de la Tomis, Pontica, 43, 2010, p. 555-62. Preda 1968 Constantin Preda, Monede coloniale rare si inedite – Histria, Callatis si Tomis – din tezaurul de la Mangalia, Studii și Cercetări de Numismatică, 4, 1968, p. 223-37. Rădulescu 1966 Adrian Rădulescu, Monumente romano-bizantine din sectorul de vest al cetăţii Tomis, Constanţa, Muzeul Regional de Arheologie Dogrobea, 1966. Rădulescu 1991 Adrian Rădulescu, Recherches archéologiques récentes dans le périmètre de la cité de Tomis, Études Byzantines et PostByzantines, II, 1991, p. 23-45. Rădulescu Adrian Rădulescu, Zidul de apărare al Tomisului, de epocă 1995-1996 târzie în reconstituirea sa actuală, Pontica, 28-29, 1995-1996, p. 83-93. 144 The roman defensive systems of Tomis Adrian Rădulescu, Constantin Scorpan, Rezultate preliminare ale săpăturilor arheologice din Tomis (Parcul Catedralei), 19711974, Pontica, 8, 1975, p. 9-54. Rykwert 1995 Joseph Rykwert, The idea of a town: The anthropology of urban form in Rome, Italy and the Ancient World, MIT Press, 1995. Scarlat 1975 Constantin Scarlat, Portul antic Tomis și peninsula submersă Tomis, o așezare geto-dacică anterioară sosirii navigatorilor greci. Cercetări de arheologie submarină, Muzeul Național, 2, 1975, p. 257-72. Soficaru 2011 Andrei Dorian Soficaru, Populaţia provinciei Scythia în perioada romano – bizantină (sf. sec. III – înc. sec. VII), Iaşi, Editura Universității “Alexandru Ioan Cuza”, 2011. Stoian 1954 Iorgu Stoian, În legătură cu decretele tomitane privitoare la paza orașului, Studii și Cercetări de Istorie Veche și Arheologie, 5, 3-4, 1954, p. 557-68. Stoian 1962 Iorgu Stoian, Tomitana. Contribuţii epigrafice la istoria cetăţii Tomis, 1962. Suceveanu Alexandru Suceveanu, Observations sur la stratigraphie des 1969 cités de la Dobrogea aux IIe-IVe siècles à la lumière des fouilles d’Histria, Dacia N. S., 13, 1969, p. 329-66. Tafrali 1930 Oreste Tafrali, Ruinele unei clădiri din epoca imperială romană la Constanţa, astăzi dispărute, Arta şi Arheologia, 5-6, 1930, p. 52-55. Teodor 2012 Alexandra Teodor, Antique Layers of Tomis, in B. Jöger, D. Comşa (eds.), ICAR 2012: (Re)writing history. International Conference on Architectural Research: Proceedings, Bucharest, 18-20 May 2012, “Ion Mincu” Publishing House, no. 307. Available online at: https://www.academia.edu/5140327/Antique_Layers_ of_Tomis Toma 2010 Natalia Toma, Tomis - Kustendje - Constanța. Topografia antică tomitană în hărţi şi însemnări de călătorie din epoca modernă (sec. XIX - începutul sec. XX), Caiete ARA, 1, 2010, p. 53-74. Zahariade 2006 Mihail Zahariade, Scythia Minor. A History of a Later Roman Province (284-681), Amsterdam, Adolf M. Hakkert, 2006. Zirra, Vlad Zirra, Petre Alexandrescu, Cercetări arheologice în Alexandrescu orașul Constanța, Materiale și Cercetări Arheologice, 4, 1957 1957, p. 88-94. Rădulescu, Scorpan 1975 145 Alexandra TEODOR Websites abbreviations111 ANCPI 2005, 2008 catheofilm.ro CCA 2000 CCA 2001 CCA 2008 cerva.ro/ tomis/ cjc.ro gzt.ro-1 2013 gzt.ro-2 2013 gzt.ro-3 2013 PUZ 2012 RAN wildwinds. com/ moushmov/ tomis 111 http://www.ancpi.ro/pages/home.php?lang=en. A web version of the orthophoto set can be seen here, http://geoportal.ancpi.ro/geoportal/viewer/index. html. Some useful instructions here: http://civil3design. wordpress.com/tag/imagery/. http://www.catheolfilm.ro/video.html http://cronica.cimec.ro/detaliu.asp?k=1144&d=Constantaorasul-romano-bizantin-poarta-mare-2000 http://cronica.cimec.ro/detaliu.asp?k=1343&d=Constantastr-Arhiepiscopiei-nr-23-2001 http://cronica.cimec.ro/detaliu.asp?k=4144&d=ConstantaStr-Mihai-Viteazu-f-n-sediu-OCPI-Constanta-2008 http://cerva.ro/tomis/ http://www.cjc.ro/MINAC/en/index.html http://gzt.ro/povestea-batranului-tomis-cu-2600-deani-sub-talpi-i/ http://gzt.ro/batranul-tomis-cu-talpile-pe-2600-de-anide-civilizatie-ii/ http://gzt.ro/povestea-unei-zile-de-arheolog-sausoldurile-rubensiene-ale-unei-venus-din-tomis/ http://www.oardobrogea.ro/wp-content/uploads/2012/ 05/8-arheologic-841x1250.pdf, Mirela Băncescu, Gh. Vecerdea, Gheorghe-Leonard Duță, „Actualizare P.U.Z. – Peninsula, Municipiul Constanța, Patrimoniu arheologic”, scale 1/2000, 2012 (2008 updated edition), pl. no 1. http://ran.cimec.ro/?lang=EN http://www.wildwinds.com/moushmov/tomis.html All websites were accessed on May 27th, 2014. 146