Skip to content
Επιθεώρηση Αδριατικής Τέχνης Search

The castellum in Mogorjelo: A Contribution to a Long-Lasting Debate

DOI 10.15291/aa.4072

Download PDF

Vol. 12 (2022)

Open access

Article metrics

  • 3 412 Abstract views
  • 1 248 PDF downloads
  • 12 Downloads per day
  • 164 Downloads this month
  • Tin Turković University of Zagreb, Faculty for Humanities and Social Sciences tturkovi@ffzg.hr Corresponding author
  • Nikolina Maraković University of Zagreb, Faculty for Humanities and Social Sciences

The late antique complex in Mogorjelo can be categorized as one of the prime Dalmatian late antique monuments. It has attracted considerable attention over the past hundred years and more, and its ground plan can be found in a number of reviews of late antique architecture in Dalmatia, as well as in the wider region. Still, although it has been studied for such a long time, there are many unanswered questions regarding this “puzzling” building complex. When a monument has such a long history of research, it is bound to become a subject of divergent and even conflicting opinions. The aim of this paper is to present the complex as objectively as possible, thus rectifying the conceptions about its form, function, and purpose in a wider organizational framework established in the province by Emperor Diocletian. The intention is not to explain each detail of the complex, but to offer a new and more comprehensive insight into the nature of the castellum itself, and to provide the scholars of late antique architecture with factography, as complete as possible, about the complex and its surroundings.

Mogorjelo, late antique architecture, Roman non-urban landscape, Dalmatia, Emperor Diocletian
Sections
  1. Abstract
  2. INTRODUCTION

INTRODUCTION

The famous complex in Mogorjelo, situated south of the town of Čapljina (2.15 km), has amazed and confused many distinguished scholars, so there has been a lot of discussion about its nature and purpose over the past hundred years. To our knowledge, Edin Veletovac was the last scholar who summarised all of the dilemmas and different conclusions that had accumulated during that time, in his excellent pa- per from 2018.1 Some of the conclusions are complementary, while others are quite contradictory; furthermore, some of them are based on selective uses of facts, or on comparisons and theories that have meanwhile been proven wrong. So, if Mogorjelo is indeed, as Veletovac called it, “the largest (late antique) architectural complex in the territory of Bosnia and Herzegovina,” it certainly deserves our attention and an- other re-examination. In order to do that, it is essential to start with some general information about the complex, and the previously published conclusions and hy- potheses about its nature, function, placement, and “specific” arrangement, in order to be able to assess all the facts, even the ones that were circumvented by former researchers, and then put them in the proper context.

THE COMPLEX IN MOGORJELO BETWEEN FACTS AND FICTION Our intention here is not to summarise and present all the details of the his- tory of research, interpretation, and reconstruction of the complex in Mogorjelo, because, as we have already said, Veletovac did it quite well. However, some of the main preoccupations should be underlined and complemented with new conclu- sions about the nature and function of the fortified complex. First and foremost, it should be noted that the hundred-year long debate has been of a much larger scope, surpassing the site of Mogorjelo itself, and is actu- ally concerned with fundamental issues regarding late antique non-urban landscape and its architecture across the Empire. It includes a number of issues concerning the organization and way of functioning of late antique non-urban landscape, the structure and management of large estates in the provinces, and the recognition of certain general principles, common across the Empire. In that respect, the discus- sion about Mogorjelo around the turn of the 4th century presents a case study (and not the only one in Dalmatia) that perfectly mirrors these wider preoccupations. The hundred year-long debate about Mogorjelo, to include a comic relief, reminds us of the introductory verses of a local folk song called “Hasanaginica”: What’s so white upon yon verdant forest? Snow perhaps it is or swans assembled? Snow would surely long ago have melted. And a flight of swans would have departed. No! not swans, not snow it is you see there, Tis the tent of Aga, Hasan Aga.2 The quoted verses illustrate quite well the dilemma about the nature and function of the late antique complex, as can be seen from the titles of some crucial papers writ- ten about it. Starting with Ejnar Dyggve's “Mogorjelo. Kastell oder Palast, 1. oder 4. Jahrhundert?” from 1958, followed by Ivo Bojanovski's “Mogorjelo – Rimsko Turres” in 1969, Noël Duval’s “Mogorjelo: villa fortifiée ou camp militaire?” from 1991, and ending with Marin Zaninović’s paper “Mogorjelo from villa to castrum.”3 Although not all of the mentioned papers end with a question mark, their content and conclu- sions imply that they actually should. Basically, all of them reflect the same dilemma: what was Mogorjelo actually? Further on, the overall significance of the issue for the general research of late antique non-urban architecture was emphasized by Dyggve at a conference in Munich in 1958: “Sofort bei der Publikation Mogorilos durch

Nikolina Maraković a Long-Lasting Debate

a b

1. Patsch erregte der Grundriss das größte Aufsehen unter den Fachgenossen, und war Mogorjelo castellum – vor allem wegen der unstreitigen Verwandtschaft des Gebäudes mit dem Diokle- (a) Satellite image of the castelllum; (b) Ground plan tianspalast. Die Forscher sahen ein, daß Mogorilo ein architekturhistorischer Beitrag published by Karlo Patsch von allergrößtem Interesse zu der damals hochaktuellen Diskussion über den Palast (Patsch /note 3/) von Split bedeutete.”4 In fact, this series of papers on Mogorjelo represent and reflect Castellum u Mogorjelu – not just the discussion about one particular site, but illustrate the general outlines of (a) Satelitski snimak kaštela; the evolution of our knowledge about late antique architecture. (b) Tlocrt koji je objavio Karlo Patsch Considering the complex itself, the discussion seems to have progressed “one step forward, two steps back.” Karlo Patsch’s publication of the Grundriss and his interpretation of the complex as one of the military forts (castella, military camps) in the line of defence Burnum-Tilurium-Mogorjelo, was actually the starting point for the ensuing discussion (Fig. 1b).5 Ejnar Dyggve’s presentation at the XI. Inter- nationaler Byzantinisten-Kongress in Munich in 1958 should be seen as a milestone in research and interpretation of Mogorjelo. However, his conclusions are very disputable. His paper was actually completely in tune with his personal vision of late antique cultural landscape, as well as the predominant perspective of German- speaking scholars.6 Dyggve’s exposition, based on his observations and works done in Mogorjelo before the Second World War, in collaboration with Rudolf Egger, certainly awakened interest in Mogorjelo once again, but the way he presented the complex is very problematic. Although Ejnar Dyggve is widely perceived (especially in Croatia) as an archae- ologist, he was actually neither an archaeologist nor an art historian by his formal education. He was an architect, and in the 1910s became a member of the student architects’ group Kanonarkitekterne, which dealt intensively with questions of archi- tectural theory.7 Although he is still occasionally praised as “the man who revealed to the Croats all the secrets of Salona” or “the Dane who unearthed the whole Adri- atic coast,”8 his conclusions about Dalmatian as well as Greek monuments have to be taken with utmost caution. He could be commended for his efforts as an enthusiast in archaeology, but not so much for his achievements in the interpretation of the monuments he studied. As Noël Duval almost bluntly stated in his “Hommage à Ejnar Dyggve et Ingrid Dyggve” from 2003, Dyggve belonged to the circle of “des architects qui étaient d'excellents dessinateurs” – as his small croquis drawings and his general “approach” were heavily imbued with preconceptions about “Gothic ar- chitecture,” mainly focused on Ravenna.9

Nikolina Maraković a Long-Lasting Debate

2. We have to agree with Duval on the context of Mogorjelo, because Dyggve re- Corinthian capitals from ally went too far into the field of speculation when it comes to the reconstruction Mogorjelo (Dyggve, Vetters / note 3/, T. VII, 3, 4) and interpretation of the complex. He wrote: “Eine Wandelhalle gegen die offene Landschaft gekehrt, ist ja ein wohlbekannter Zug der freigelegenen Paläste spätan- Korintski kapiteli iz Mogorjela tiker Zeit. Da kein einziges Fragment von einem Säulenportikus an dieser ausge- dehnten Seite bei der Grabung aufgefunden worden ist, habe ich mir die Fassade nicht mit Säulen, sondern mit gemauerten Pfeilern und Bogen vorgestellt.”10 So, Dyggve created his own image of the façade, in spite of the fact that Hermann Vet- ters, for example, wrote in the same publication that the porticoes in front of the main building in Mogorjelo were formed with 14 columns at the front side, noting even the distance between them.11 He also noted that two bases had been pre- served, as well as a few capitals, which, at least judging from the old photographs, indicate Diocletianic provenance (Fig. 2).12 Thus, it seems that Dyggve disregarded some finds. He himself wrote about his approach: “Wie es aber geht, wenn man eine Bauruine an Ort und Stelle analysiert: neue, erklärende Einzelheiten erschein- en. Sobald ich den Generalplan aufgezeichnet hatte, und die Teilung in Wirtschaft- shof und Knechthof sich klar auffassen lies, wie auch die Erklärung des langen Korridors als Wandelhalle begründet war, musste die Funktion des Gebäudes als Kastell bezweifelt werden. Ich hatte meine Auffassung dahin zu ändern, das Mo- gorilo ein Herrenhaus eines Latifundiums in dieser überaus fruchtbaren Gegend gewesen ist”13 – confirming in a way Duval’s “architects qui étaient d’excellents dessinateurs”?14 He did not notice that the walls of an oil press and other struc- tures in the north-west quadrant of the complex were, according to stratigraphical data, later than the rest of the complex.15 It is also quite indicative that Dyggve mentioned that Rudolf Egger, an experienced professional archaeologist, disagreed with his new interpretation. So, Dyggve concluded that Mogorjelo was a fortified manor house on a lati- fundium, in an exceedingly fertile region, reaching for comparisons with buildings depicted in North African mosaics, as well as with Diocletian’s palace in Split and the complex in the Polače bay on the island of Mljet.16 His ground plan and his re- construction show considerable arbitrariness, which is quite obvious from the com- parison between his ground plan from 1931 and those published in 1966 (Fig. 3).17 Dyggve even disregarded the preserved remains of the walls of the main building. Further on, he attributed certain functions to the three central spaces, labelling them as triclinium, vestibulum, and aula, and he also drew some non-existent openings, elongated certain walls, included three-flight staircases in some of the towers, etc. However, the most problematic is his reconstruction of the elevation of the complex:

Nikolina Maraković a Long-Lasting Debate

above the porticoes surrounding the courtyard, which he reconstructed with disre- gard for the actual stratigraphy, he drew the first floor, although there were no finds to corroborate such an imaginative reconstruction.18 Unfortunately, Dyggve’s reconstruction of the complex, however problematic, remained widely accepted for a long time. One of the researchers who elaborated on his ideas and felt obliged to round up Dyggve’s “insights” was Hermann Vetters, who published them in an extended version in 1966 (Fig. 3b). Vetters’ contribution was actually pretty much in accordance with the Austrian tradition of precise and objective documentation of archaeological finds, so the chapters dealing with the measurements and descriptions of the finds in various sectors are quite meticulous and instructive. Besides, it appears that Vetters had a keen eye for measurements and ratios, as can be seen from his observations on the dimensions and propor- tions of the main building and the southwest corridor.19 He also had an eye for details, and he noted even the smallest of them, such as the postholes for wooden beams at the height of 2 m in the southwest corridor, which must have supported the wooden ceiling of the portico.20 In spite of his careful analysis of the remains of the complex, Vetters eventually adopted Dyggve’s preconceptions and elaborat- ed them further. Unfortunately, his interpretation went even one step back, as he concluded that the main building looked like a Porticus Villa mit Eckrisaliten, and thus could not have been a military facility.21 With that preconception, he went on 3. to repeat and elaborate on Dyggve’s conclusions, contradicting himself at certain Dyggve’s ground plans from points,22 most evidently when it comes to the reconstruction of the elevation of 1931 – (a) ground plan with the the main building.23 Vetters’ contradictory conclusions must have been the result assumed arrangement of rooms; (b) detailed ground plan of all of being torn between Dyggve’s imagination and the reality of documented finds. of the documented remains He concluded that there was a balcony on the first floor, and that the arrangement published in 1966 (Dyggve, Vetters of the rooms on that floor was analogous to the one on the ground floor, reaching /note 3/, Tafel XIX, Beil I) again for parallels with Diocletian’s palace in Split. However, it seems completely Dyggveovi tlocrti iz 1931 – (a) tlocrt unconvincing that relatively frail columns and, in some instances, only 0.66 m s pretpostavljenim rasporedom thick partitioning walls could have supported all these rooms (the vestibule, sa- prostorija; (b) detaljan tlocrt sa svim dokumentiranim nalazima, lons, etc.) on the first floor; and that is why Vetters reached for Dyggve’s assump- objavljen 1966. g. (fotografija tion that the first floor must have been supported by pillars, although the finds autora) clearly implied otherwise.

a b

Nikolina Maraković a Long-Lasting Debate

Eventually, neither Dyggve nor Vetters drew a reconstruction of the façade of the main building. Dyggve made many sketches of the elevation from various stand- points, but not a single one with a view of that façade (Fig. 4). On the other hand, all of his sketches of the complex were “endowed” with a number of features quite illogically borrowed from Diocletian’s palace in Split. So, although Vetters’ analysis brought a lot of new information, his conclusions remained on the heels of Dyggve, no matter how contradictory in themselves they were. Just about the same time, Christian churches in Mogorjelo attracted the atten- tion of Đuro Basler, who presented his reconstruction of the whole complex in 1972, with a view of the façade of the main building (Figs. 5, 6a).24 Expectedly, he visual- ised what he could learn from his only source – Dyggve’s and Vetters’ booklet from 1966,25 and thus created his well known and most often reproduced reconstruction. In Basler’s reconstruction, the ground floor of the main building is opened towards the courtyard with arched pillars, and the upper floor with columns. He also made some alterations and improvements to Dyggve’s reconstruction – for example, above the porticoes he imagined a slanting, not a flat roof, although such an alteration did not make sense from the point of defensive function of the fortification, because it did not leave free space for a walkaway around the building. He eliminated some of the similarities with Diocletian’s palace in Split, such as the large arched openings in the outer walls of the complex. A more significant improvement was the removal 4. of Dyggve’s division of the courtyard into two areas: a farmyard (basse-cour) on the Dyggve’s sketches eastern, and a formal forecourt (cour d’honneur) on the western side.26 However, in (reconstructions) of the castellum from 1931 (Dyggve, Vetters /note this last respect, Basler’s reconstruction contradicted his conclusions that accompa- 3/, Abb. 24, 25) nied the illustration. He made an even bolder claim than Dyggve and Vetters – that Dyggveovi crteži (rekonstrukcije) the remains of an earlier structure, essentially economic in nature, were partly pre- kaštela iz 1931. served, restored, and enveloped by the complex.27 Again, Noël Duval emphasized

Nikolina Maraković a Long-Lasting Debate

5. Basler’s reconstruction of the complex from 1972 (Basler /note 24/, fig. 9) Baslerova rekonstrukcija sklopa iz 1972.

that such an interpretation was not accurate, and in fact represented “une grossière erreur d’interprétation”.28 Basler labelled the main building “grand palace” and used the term villa for the whole complex.29 This terminology is rather confusing, and it seems as if the terms were used without properly distinguishing their meaning.30 Finally, and this is quite important for our discussion, Basler concluded that the villa served to supply the town of Narona, although he did not specify what kind of supplies that would be.31 More or less about the same time, in 1969, Ivo Bojanovski presented yet another interpretation of the site. In many respects, Bojanovski’s conclusions were a significant advance in understanding the form and function of the complex. He explicitly con- cluded that the position of the complex and its castrum-like form, with an obvious de- fensive function, indicate the original function of the whole complex. 32 He contested the core of Dyggve’s and Vetters’ logic, and presented a convincing critique of their ideas about Mogorjelo as a palatium or palatiolum, as well as of their dependency on comparisons with Diocletian’s palace in Split. He contrasted the outlines of the com- plex in Mogorjelo with the grand complex in Višići, just 3.25 km south of Mogorjelo.

a

6. Basler’s reconstruction of the b complex – (a) façade of the main building; (b) northern front of the fortifications with the main gate (Basler /note 24/, figs. 10, 11) Baslerova rekonstrukcija sklopa – (a) pročelje glavne građevine; (b) sjeverna strana fortifikacija s glavnim ulazom

Nikolina Maraković a Long-Lasting Debate

The complex in Višići, a cluster of buildings that developed over the centuries, was of a quite different character and not fortified. Bojanovski accurately noticed the dis- crepancy in form and function of the two interrelated complexes, as will be explained in the following text. There is another crucial difference between the two. Namely, as Basler himself noticed in 1965, and as Bojanovski restated, the complex in Mogorjelo was deprived of extensive luxury, and even of some basic installations required for everyday civilian life – baths, heating, etc.33 So, Bojanovski’s observation that the whole complex exuded “military simplicity” is, indeed, well founded, as well as his conclu- sion that the complex was certainly not a praetorium fundi.34 Bojanovski’s other conclusions are less convincing and not so well grounded. For example, his dating of the complex in the second half of the 3rd century, at the time of Emperor Gallienus or Emperor Aurelianus, can be easily refuted, as there is enough evidence of the “Diocletianic” style, especially in the execution of capitals, but even in the form of the castellum itself. In that respect, all of his predecessors were right.35 As for his reconsiderations of the position of the complex, it seems that he was right, although he never offered a full explanation for such a position of this kind of castellum. It is rather strange that Bojanovski changed his opinion with time, as in 1988 he stated: “…later research did not entirely confirm Patsch’s thesis, because it turned out that what was preserved was a 4th-century structure, whose function was by some interpreted as a fort (i. e. the same as Patsch) and by others as the centre of a sizeable estate, i.e. as a large villa rustica or palatium, which Mogorjelo indeed became, but in late antiquity.”36 This presents an important and quite incomprehensible diversion from his previous conclusions and, despite a more accurate dating, was a step back in the interpretation of the complex. The next and crucial milestone in the history of research of the site was Noël Du- val’s paper from 1991.37 In the most structured manner, in three short chapters, he dis- mantled all of the conclusions made by Dyygve and Vetters, and brought the discus- sion about Mogorjelo back on track. He contested almost all of Dyggve’s conclusions, regretting that Vetters had not restored the truth about Mogorjelo finishing off with Dyggve’s theories.38 Regardless of Duval’s style when writing about Dyggve, in which there may have even been something personal, he was to the point. Imagination, ar- bitrariness, ad hoc comparisons based on a rather limited knowledge of comparative material, and disregard for some archaeologically confirmed facts make Dyggve’s con- clusions very disputable.39 Duval also concluded that the cohors quingenaria equitata of Karlo Patsch made all the sense in this kind of structure, and that quite a different im- age of the complex emerges when cleansed from Dyggve’s unfounded speculations.40 Finally, we come to the one of the latest interpretations of the complex, published by Marin Zaninović in 2002.41 Except for repeating the history of research, which makes the greatest part of the paper, his conclusions are in accordance with those of Bojanovski in 1988. According to Zaninović, the site evolved from a fertile re- gion villa to a late antique castrum, and that happened because of the threats in the region. However, he did not offer a precise date of that transformation. Instead, he actually repeated an old hypothesis, just using the term castrum instead of the usual Dyggvean terms.42 With this we bring our introduction to a conclusion, hoping that we have pro- vided satisfactory insight into the past discussion. The result might be compared to the typically Mediterranean bocce (Dalm. balote) game. The opposing sides that par- ticipated in the history of research of Mogorjelo made some quite interesting moves, and the “team Dyggve” seems to have made more points, but received a couple of voli, so is still far away from the pallino. As Veletovac demonstrated,43 the game is not over yet, and nobody scored full 15 points.

Nikolina Maraković a Long-Lasting Debate

NEGLECTED ISSUES AND UNANSWERED QUESTIONS The whole century of discussion about Mogorjelo has left us with a number of unresolved issues. What was Mogorjelo actually and which term would best describe its function? To which category of late antique buildings does it belong? What ex- actly did it look like? Mogorjelo was a fortified site, but we are still wondering about the reasons for its defensive function, because at the end of the 3rd and the beginning of the 4th century, there were no external threats to this or any part of coastal Dalmatia. Still, the building had thick walls, eleven three-story towers, and was in all probability provided with modern military equipment, such as the 4th-century onagra and its ammunition.44 That kind of military equipment implies that the complex could not have been a pseudo-fortified palatium like Diocletian’s palace in Split, but an actual fortification. Further on, there is the question of its position. Although some of the researchers, in the first place Bojanovski, made an effort to locate Mogorjelo inside the network of Roman roads, stations, and settlements, none of them explained why the fortification had been built on that particular spot around the turn of the 4th century. As some of the scholars have already mentioned, Mogorjelo is at only 7 km air distance from ancient Narona, and until the 19th century it was in the middle of marshes and swamps, in an area plagued by malaria and always suffering from inun- dation and hunger.45 That it was situated in a malaria-stricken, swampy environment can also be deduced from Patsch’s project of protecting the site by planting cypress trees.46 Still, this does not mean that there was no melioration in the valley in ancient times, as the prosperous estate in the neighbouring Višići testifies. In the case of Mogorjelo, considering the more than probable existence of a villa on the site (with or without an oil-press), as well as 164 agricultural tools and implements found on the site, clearly testify that Mogorjelo and its surroundings were originally suitable for agricultural exploitation.47 Further on, it is strange that nobody noticed the fact that Mogorjelo, overlooking the river of Neretva, is just 1 km north of the place called Struge, south of which the Neretva becomes sailable and from where the delta begins. Just south of Struge, the Neretva widens to about 100 m, and from that point becomes sailable even for the largest of ancient ships. If we consider all the abovementioned, it seems that the castellum was actually there for several reasons. It held a highly defendable position, as it was located on a small hilltop in the middle of the marshes. It oversaw the Neretva from one of its strongest and probably the best equipped towers (the round one), and had an excellent view across the open plains. Finally, it was just 1 km from the last stop for any large vessel moving up and down the sailable part of the river. All that points to the fact that Mogorjelo was overlooking and defending something, but not Narona. Also, it could not have been the supply centre for Narona, as Basler assumed. So, the question remains: what was so important and valuable to be de- fended in this way? We will return to this issue in the following passages. Considering the position of the complex, it is quite strange that none of the re- searchers examined the reason for the orientation of the fortress to the northeast. In fact, the majority of published ground plans can be deceiving, as they are printed ei- ther upside down or just incorrectly.48 The main gate was on the north-eastern side.49 It should also be noted that the main entrance was accentuated by two ortosthates, one decorated with the motif of a vine tendril with grapes, and the other with an interesting depiction of a whole stem of acanthus mollis.50 Both of these are signs of abundance and plenty. Furthermore, the main building, with its large central room, was oriented in the same direction. Thus, it seems very probable that somebody, or something, was expected to arrive from the north, and not from the direction of Na-

Nikolina Maraković a Long-Lasting Debate

a b

7. rona. That something was leaving the complex through the side gates, defended by The complex in Višići (Kućišta) – towers; and an escape exit was left in the back, southern side of the complex, which (a) Reconstruction of the complex by Busuladžić (Busuladžić, /note was also the least defended. That “emergency exit” was connected with the central 44/ T. 2a); (b) Ground plan of the room of the main building through a passage in one of the lateral rooms (Fig. 3b). complex in Višići (Čremošnik / With such an exit, there was no real threat to the complex. But, what was so valuable note 63/) that it had to be so carefully defended, and was at that time coming from the north? Kompleks u Višićima (Kućišta) – When discussing the surroundings of Mogorjelo, Bojanovski introduced the es- (a) Busuladžićeva rekonstrukcija tate in Višići, although he did not discuss the relationship between the two com- kompleksa; (b) Tlocrt kompleksa u Višićima plexes. They were actually complete opposites in form and function. As opposed to Mogorjelo, there are no fortifications in Višići. However, in Višići there are plenty of utilitarian facilities, which are almost non-existent in Mogorjelo (Fig. 7).51 Višići was over time equipped with at least four balnea, hospitia, taverns, blacksmith’s premises, etc., as opposed to Mogorjelo, in which there was no such luxury. And Višići was just 3.25 km south of Mogorjelo, on the other side of the Neretva, approximately 35 minutes of light walk. We shall return to Višići and its relation to Mogorjelo in the following section. After explaining the position and orientation of the complex in Mogorjelo, there are also a number of things that should be said about its form and function, which was not properly explained in the past research. Basler and Bojanovski only superfi- cially noticed that the complex gave an impression of military sternness and simplic- ity, but a thorough analysis of the ground plan or the relations between the structures adjacent to the main building has never been made. And although Dyygve tried to identify the functions of certain rooms in the main building, he just made a lot of rather unclear conjectures (Fig. 3a). In our opinion, it is evident that the central part of the main building could not have been a basement, storage, or anything of the kind. On both sides of a somewhat larger central room, there are spacious rooms of more or less the same size, all with entrances from the porch. Most of them com- municated with each other, except maybe for the two rooms most westwards. These rooms do not look like living spaces, but rather like spacious, interconnected offices. Then, there are quite obvious differences in the thickness of walls within the main building: the walls of the central room and the corner pavilions are almost twice as thick as the majority of the partitioning walls; so, the logical conclusion would be that these parts of the building were higher than the rest. In our opinion, there was

Nikolina Maraković a Long-Lasting Debate

no continuous second floor, as it has been proposed in previous research, and the staircase preserved in the south-eastern corner probably led to the upper floor of the corner pavilions. As it is visible from the ground plan, these pavilions did not com- municate with the central part of the building, but probably functioned as separate entities, with their own entrances. They actually look like isolated storages to which the entrance was strictly controlled (Fig. 3b). The first room to the east of the central room opens to the south through a back door. Thus, it could have provided access to the abovementioned back exit from the complex, as well as to the back corridor, which ensured moving freely along the back of the building. As Bojanovski suggested, the lateral corridors in the southern part of the complex, at both sides of the main building, were separated by doors from the courtyard (which is why there are protruding fragments of walls in front of both lateral corridors, as can be seen from the ground plans). So, the cubicula in the south-western quadrant of the complex could be entered only through the closed lateral corridor, and were thus for some reason probably separated from the rest of the complex. To what purpose this part of the complex might have served is a ques- tion that still needs to be answered.

MOGORJELO – NOTHING OF THE PREVIOUSLY PROPOSED? Having said all the above, we should now try to answer the questions we have posed. First, we shall examine the position and orientation of Mogorjelo. As we have explained, its position was quite particular, with the main entrance on the north- ern (more precisely, north-eastern) side. Bojanovski was probably right when he identified Mogorjelo as a late antique Ad Turres, with convincing arguments,52 and Busuladžić has demonstrated that there was a military detachment in Mogorjelo at the time when the castellum was built.53 As we have mentioned above, at the time when the castellum was erected there was no outside threat in the immediate sur- rounding, and the only threat inside the province could have come from banditry and looting on military routes, as a consequence of the long-lasting 3rd-century crisis.54 And this is the point where the cohors quingenaria equitata emphasized by Patsch and Duval becomes relevant. There may have been a cohors stationed in Mogorjelo, as these were precisely the kind of units that Diocletian used on various posts along the limes and elsewhere, or there could have been smaller cavalry units (turmae). A mobile unit made of a few squadrons is certainly to be expected on site.55 So, we can conclude that the military was stationed in Mogorjelo to protect something that was brought from the north. The most valuable commodity coming from the north-eastern part of the Dalmatian hinterland was – income from the taxes or valuable ores. One would assume those kind of goods were transferred by a land route, on a road connecting Salona with the eastern Dalmatian hinterland, consequently sometimes referred to as the “mining road”.56 Although this may have been true for some stable times, at the end of the 3rd century it could hardly be ex- pected that the valuable imperial cargo should be transferred 400 km towards Salona by a dangerous land route.57 In fact, the existence of a complex like Mogorjelo, with its “treasuries”, fortifica- tions and military crew, confirms that such a route did exist, and was formed by the end of the perilous 3rd century. Even if it was not exactly the one we proposed here, Mogorjelo may have been the final station for the cargo arriving from the eastern part of the Dalmatian deep hinterland.58 That the Neretva delta had always been the final destination for the transport of valuable ores from the north-eastern part of the Dal- matian hinterland, and that this route was the most logical for transport, especially of gold and silver, is confirmed by recent investments in the revitalisation of the gold

Nikolina Maraković a Long-Lasting Debate

8. Topographical map with indicated major roads, bridges, and settlements south of Mogorjelo (Yugoslav People’s Army map, 1:25000, alterations by T. Turković according to Bojanovski, Sergejevski, Patsch and Ballif) Topografska karta s označenim glavnim cestama, mostovima i naseljima južno od Mogorjela

and silver mine in Vareš, north of modern Sarajevo. (This huge investment is current- ly worth about 1 billion US dollars, but it will be more than profitable for the investor, as it is estimated that solely at the two activated excavation sites there is 1.01 g of gold per ton of excavated material, and 120 g of silver. Next to those, there are much larger deposits of zinc, lead, copper, baryte, etc.)59 As has been announced, all of the cargo will be transferred via Sarajavo to Mostar, and from there to the harbour of Ploče (Croatia) on the Adriatic coast.60 So, in the final stages of transport, the same route will be used as the one that was used in late antiquity. (Even the route between Mo- star and Čapljina will be used, although the modern route will follow the A1 highway from Čapljina and then continue westwards, instead of proceeding southwards.) With the announced repairs of the old Sarajevo-Ploče railroad, the cargo will pass through Mogorjelo, precisely 426 m westwards from the late antique fort, towards Metković (Fig. 8). So, it seems that the basic logic of transport has remained the same over the centuries – by the fastest, most secure, and cheapest possible route to the Adriatic. The final destination for valuable cargo coming from north-eastern Dalmatia (today’s eastern and central Bosnia) has always been the Neretva delta, and any cargo carried to that destination had to pass through Mogorjelo. If silver was indeed the precious cargo transferred via Mogorjelo, it becomes clear why the castellum was built at that exact spot. Easily defendable, with a clear view over the open plain, equipped with a cavalry detachment that could escort the transport in its last stage, and with treasuries in which silver could be kept secure, waiting to be shipped to the other side of the Adriatic when the time came for the annual round of collection of imperial goods. From where exactly the goods were shipped away is a question for future considerations. Narona is less likely, because its harbour could never receive larger river vessels,61 but there may have been a harbour somewhere around Višići, where the Neretva was always wide enough for even the largest boats to sail and manoeuvre freely.62 So, we suppose that the position and orientation of the castellum in Mogorjelo sug- gest that it was a collection point for valuable cargo arriving from the north. We may add that the Mogorjelo fort was at the same time, or exclusively, a collection point for the local taxes, which were stored in it until the income was shipped further. This is-

Nikolina Maraković a Long-Lasting Debate

sue remains open for further considerations. Either way, the imperial officials would then operate and manage the arriving cargoes, counting the input and noting the size of the future shipment. The layout of the building complex substantiates the same con- clusion. The arrangement of rooms in the main building suggests its function as an office building, with a large central room that was probably intended for the formal reception of arriving parties by some of the imperial officials. We can further speculate about the presiding official, who may have been in the service of Rationalis summarum Pannoniae secundae, Dalmatiae et Saviae, Procurator rei privatae per Dalmatiam, or Praepositus thesaurorum Salonitanorum, Dalmatiae, or one of the three officials them- selves, receiving reports on the arrival of cargo, and reports from the clerks and no- taries responsible for receiving, counting, and preparing of cargoes for shipment. The general outline of the building, quite surprisingly, resembles the so-called “curia” in Domavia, excavated by Vaclav Radimsky.63 In both cases, there is a central area, which Radimsky called a tribunal, and lateral rooms for the clerks, as well as additional lat- eral rooms for storage of the same kind of goods – silver or other valuable goods. Counting and noting the incoming load was a painstaking work for several clerks, so the administrative staff of Mogorjelo must have consisted of a number of clerks that worked in the main building, and may have resided in cubicula in the south-western quadrant, separated from the military personnel. The picture of the complex in Mogorjelo becomes much clearer if we try to draw a reconstruction based on actual archaeological finds – we have to emphasize that we intentionally left some of the details of the upmost part of the elevation unfinished, as they can hardly be hypothesized from the archaeological remains (Figs. 9, 10). However, the formal and stern character of the main building clearly emerges, sug- gesting its official status. Instead of the previously imagined romantic piano nobile at the first floor, we see a series of ground floor offices and two lateral “treasuries”. The 9. general impression of the main building and its U-shape might recall a Zentralhof in New proposed reconstruction of the complex, total (T. Turković, the form of a Villa mit Eckrisaliten, on a Streuhof-like estate, but actually there are a 2022) number of differences between the two.64 The outline of the main building could also be compared to various other structures built during Diocletian’s reign; for example, Novi prijedlog rekonstrukcije kompleksa, total the stern impression of the building with an accentuated central room and lateral

Nikolina Maraković a Long-Lasting Debate

a b

c d

10. offices might bring into mind Diocletian’s principium with a “Temple des Enseignes” Proposed reconstruction of in his camp in Palmyra (Fig. 11).65 In spite of apparent differences between the two, individual parts of the complex – (a) View from the north-eastern the logic of spatial arrangement is the same. Moreover, the complex in Mogorjelo tower; (b) View from the east; (c) as a whole may be pretty convincingly compared to the central part of “Diocletian’s View through the portico of the camp” in Palmyra, with its parade square surrounded by cubiculi.66 So, the square in main building, from the west; (d) front of the main building in Mogorjelo could be also seen as the parade square for Rear exit (T. Turković, 2022) military detachment stationed in the fortress. Prijedlozi rekonstrukcije pojedinih The dating of the castellum in Mogorjelo is quite unquestionable, given that the dijelova kompleksa – (a) pogled sa sjeveroistočnog tornja; (b) Pogled shape of the complex, its organisational logic, and its building technique and decora- s istočne strane; (c) Pogled kroz tion, are of Diocletianic origin. It is a classical so-called “Diocletianic type” of fortifi- trijem glavne građevine, sa zapadne cation, with a Diocletianic quadriburgium ground plan.67 As Lander has noted, these strane; (d) Stražnji izlaz forts look neither as “early” Roman fortifications, nor like Byzantine fortifications.68 They are a type of their own. But, although it can be inferred that Lander defines them as a separate type, Michel Reddé has argued that they represent an evolution in relation to models inherited from the Principate, rather than a sharp break.69 In many cases, Diocletian and his co-rulers indeed restored and updated the earlier forts, but there are distinctive elements which make these forts distinctively Diocletianic or Tetrarchic – the protruding towers, the multiplication of square towers, the transfor- mation of principia, etc. Examples of this type are numerous, and so are the analogies for Mogorjelo.70 In spite of minor divergence in internal organisation, this type follows the same scheme as Mogorjelo. Lander presented just the examples that are attested beyond doubt by inscriptions as Diocletianic. Among the examples, forts like those in Aquae Herculis, Qasr Qarun, M’doukal (Centenarium Aqua Viva), Qasr Bshir, or Deir el-Kahf are undoubtedly based on the same architectural concept.71 However, there are many more forts, rearranged at later times, which show Tetrarchic features. An interesting example is the large fort in Zeiselmauer (Cannabiaca?) whose principium

Nikolina Maraković a Long-Lasting Debate

11. was structured almost in the same way as in Mogorjelo.72 Though the towers were ob- Ground plan of the principium viously reshaped in Constantine’s time, the fort is undoubtedly of Tetrarchic origin.73 in “Diocletian’s camp” in Palmyra (Gawlikowski /note 61/) Along the Syrian limes there are numerous examples of Tetrarchic/Diocletianic quad- riburgia with the same distinctive features, recognisable even though these forts were Tlocrt principija u “Dioklecijanovom kampu” u Palmiri restructured in later phases. For this paper, especially interesting forts are those in Qasr el-Hallabat and Qasr el-Azraq, as they are closely related to those in Deir el-Khaf and Qasr Bshir, the later one being formally a close relative to Mogorjelo.74 Of course, there were simpler Diocletianic ground plans for forts than the one in Mogorjelo, but all of them were qudriburgia with protruding towers and a distinctive organisation of the inner grounds, as in such important places as Yotvata (Ad dianam)75 or Mezad Tamar (Tamara?), where a quadriburgium-type Aurelianic fortification (38 × 38 m) was complemented by distinctively protruding Diocletianic square towers.76 Among the smaller Diocletianic castella, evidently related in conception, especially interesting is the group around M’doukal (Centenarium Aqua Viva, Algeria), where an inscrip- tion precisely dates the fort into the time of Diocletian and Maximinian.77 This 86.80 × 85.90 m fort is obviously a representative of the North African type of Diocleti- anic forts, such as those in Seba Mgta and Bourada.78 Although they belong to the same type of fortification, the fort in Bourada, with its distinctive U-shape of the main building, reminds again of Mogorjelo, not to mention the arrangement of the barracks along the perimeter, or the main gate with a propuganculum and protruding square towers surrounding the gate.79 All of the major features found in these castella are also present in Mogorjelo, even in a more developed form. The only part which might not be from the time of Diocletian and the Tetrarchy is the round tower, as it does not conform with Diocletianic structures. All these comparisons point to the conclusion that Mogorjelo was part of Dio- cletian’s programme of reorganisation and optimisation in gathering the crucial re- sources in Dalmatia, as well as across the Empire. Building such a castellum with such a purpose would have been in accordance with his intent to upgrade and secure the efficiency of using all the available resources in order to counteract the economic crisis in the Empire. Mogorjelo could never have functioned as a private fortified manor house on a latifundium, as Dyggve concluded, nor could its main building be called a palatium. In that respect, Duval was quite right. Also, Mogorjelo could not have been an economic centre supplying Narona, as Basler has suggested, and it was not a villa rustica either, so this term should be avoided in further discussions.

Nikolina Maraković a Long-Lasting Debate

How the castellum functioned in its surrounding is also an interesting matter, il- lustrating the most probable events at the end of the 3rd century. As both Basler and Bojanovski noted, and we have mentioned above, the complex was ill equipped with utilitarian services and facilities for regular civilian life, such as blacksmithery, baths, shops, taverns, stables, etc.80 Nevertheless, all that could be found in the vicinity, in the complex in Višići. The history of Višići (Fig. 7) goes back to the late 1st century, when the site was obviously occupied by an estate of a wealthy proprietor, as can be deduced from the ground plan of the main and oldest edifice (B). Its dating is corroborated by the front wall, made in the technique of opus reticulatum. However, some time during the 2nd or early 3rd century, something changed in the system of management of the estate, supposedly the conductor or the owner, which was followed by considerable transformations in the main building. It may be presumed that this happened during the first decades of the 2nd century, as a number of numismatic finds from this time were found on the site.81 To the north-east of the main building, an adjacent facility for the production of bricks, tiles, and possibly ceramics (A) was added, and about the same time the main building was decorated with lavish mosaics. It was still used as a reception area, as may be concluded from numerous remains of shells (oysters, spiny oysters, Saint James, les praires, murex brandaris, etc.), discarded outside of the court- yards, around the building B. Irma Čremošnik was inclined to date many of the lavish mosaics in the central building to this time; probably at the same time the hypocaust was installed, and the western wing of the main building was transformed into small baths with exedrae added to the existing rooms. It is not quite clear when the building C was added to the complex, but it could not have been before the construction of the building A and the rearrangement of the main building. Building C contained another set of baths, which must have been intended for a different kind of users than those who visited the main building. Next to the baths, around the central courtyard of the building C, there were chambers that could be interpreted as shops or tabernae by the numerous animal remains found there. Some of its structures look like stables, and in the courtyard, there was a well. So, it seems that this building was intended for more common guests. Finally, there is the building D, the third one whose dating could be established with considerable certainty. The remains of its architectural decoration could be dated to the end of the 3rd or the beginning of the 4th century, so it should be contemporary with Mogorjelo, although it could be also dated a few decades later.82 Most of the building was heated, some of its parts look like a hospitium, and in some other parts baths can be recognized, the third set in the complex.83 A lot more could be said about the complex in Višići. However, the fact that, at least at the end of the 3rd century, the complex was not a private estate living on agriculture, is crucial for the emerging picture of the transformation of Dalmatian non-urban landscape between antiquity and late antiquity. The best guess would be that the estate had found its way into imperial hands already by the middle of the 2nd century, when the system of management changed, and when also the diversi- fication of services and the offered or produced goods occurred.84 And although the management system, as well as the conductor of this multifunctional complex, changed, it must have been managed by somebody who gathered significant income for the imperial fiscus, and lived a lavish lifestyle. Finally, the income must have been especially high at the time of Constantine the Great, since the second, equally big lot of coins found at the site belong to that time. Judging by the numismatic finds, the good times came to an end during the second half of the 4th century, or to be more precise, at the time of Emperor Gratianus.85 In our opinion, Višići and its conductor must have had a steady and growing clientele, otherwise such a complex would be meaningless. The clientele in need of such services and commodities must have been stationed at, or arriving in Mogo-

Nikolina Maraković a Long-Lasting Debate

rjelo, since Mogorjelo lacked all such facilities. The finds in Višići show a progres- sive growth of its size and possibilities, and it seems to have reached the peak of its capacities exactly at the time when Mogorjelo was built. At that time, the new part of the complex in Višići was decorated with new mosaics, and the walls of the main building were dressed in real marble, covering the frescoes that only imitated this kind of extravagant decoration.86 It would be hard to imagine that at the end of the 3rd century both Mogorjelo and Višići were anything else but imperial estates. When exactly they became impe- rial property is open for debate, but judging by the history of Višići, that might have happened already in the 2nd century. Among others, Elio Lo Cascio has most clearly presented the mechanisms of enlargement of imperial properties over the centuries, and a logical conclusion would be that the province of Dalmatia shared the same fate as the rest of the imperial provinces.87

CONCLUSION Significantly more could be said about the history of research, and about the development of the complex in Mogorjelo. In this paper, we were concerned neither with the early history of the site, nor with the subsequent development of the com- plex following the erection of the castellum. Our intention was to differentiate facts from fiction, and to provide the most objective and most logical interpretation and reconstruction of the complex at the time when it was built, including both its ap- pearance and its purpose. Trying to revise all of the previous imaginary and unsubstantiated interpreta- tions, and taking into account all of the known facts, we have come to the conclu- sion that the castellum in Mogorjelo was neither a palace nor a villa, or simply a military installation, but an administrative centre built for a specific purpose – to serve as the collection point for valuable imperial commodities arriving from the eastern part of the Dalmatian hinterland. As such, its construction was perfectly in accordance with Diocletian’s policy of optimisation of the system of harvesting all the available resources in order to push the Empire through the economic crisis of his time. In that respect, those who saw the “palace” in Split as a direct comparison to Mogorjelo were not entirely wrong, as the future Diocletian’s retirement residence seems to have been built as a complex with similar function, although a much larger one.88 However, considering certain formal features and specific internal organisa- tional logic, the comparisons are farfetched. It could only be concluded that Mogo- rjelo and the first complex in Split were compatible in general form and function, and belonged to the same general system of administration, control, overview and collating of imperial goods inaugurated by Diocletian. That system depended on the strict control of production, import, export, and transport of the imperial goods by imperial officials, supported by the military personnel. There are at least two other examples of the same system in coastal Dalmatia, one related to the quarrying and the other to the collection of portorium. The first one was on the island of Brač, with its centre in Mirje, and the second one on the island of Mljet, with its centre in the Polače bay.89 Those were all physical manifestations of Diocletian’s economic reforms, which were implemented in a way the Emperor him- self was best accustomed to – the military way. Thus, it should not be a surprise that the generation of Dyggve intuitively recognised some resemblances between Mogo- rjelo, Split, and Polače, although they went too far and astray with their interpreta- tions. Still, Mogorjelo, as an administrative centre and a collection point, backed up by a military crew, fits perfectly into that system and in Diocletian’s mindset. So, Mogorjelo indeed belonged to the late antique equivalent of Hasan Aga (in a broad sense: good master) – Emperor Diocletian.

Nikolina Maraković a Long-Lasting Debate

NOTES 1 EDIN VELETOVAC, Osvrt na arhitektonski kompleks na Mogo- of the Danish National Museum in Salona, Dalmatia, ​​ financed rjelu kao problem historiografije XIX, XX i XXI stoljeća, Acta by the state Rask-Ørsted Fondet. Illyrica. Godišnjak Udruženja BATHINVS II/2, Sarajevo, (2018), 8 See the praises in Croatian daily press: Večernji list, August 24, 317-334. 2014, or Jutarnji list, October 14, 2014. 2 “Što se bijeli u gori zelenoj? / Al su snijezi, al su labudovi? / Da su 9 NOËL DUVAL, Hommage à Ejnar Dyggve et Ingrid Dyggve. La snijezi, već bi okopnuli / labudovi već bi poletjeli. / Nit su snijezi, théorie du palais du Bas-Empire et les fouilles de Thessalonique, nit su labudovi / nego šator age Hasan age.” ALIJA ISAKOVIĆ, Antiquité Tardive 11, (2003), 273-300. Hasanaginica: 1774-1974: prepjevi, varijante, studije, bibliografi- 10 EJNAR DYGGVE, HERMANN VETTERS (note 3), 13. ja, Svjetlost, Sarajevo, 1975, 288 (trans. William Edmondstoune Aytoun, 1844). 11 In Vetters’ analysis of the “small square” in front of the main build- ing. See EJNAR DYGGVE, HERMANN VETTERS (note 3), 19. 3 EJNAR DYGGVE, Mogorjelo. Kastell oder Palast, 1. oder 4. Jah- rhundert?, reprinted as chapter II in EJNAR DYGGVE, HER- 12 The two capitals were associated with the site for a good rea- MANN VETTERS, Mogorjelo. Ein spätantiker Herrensitz im son (the bases of the columns were found in front of the main Römischen Dalmatien, Schriften der Balkan Komission, Wien- building), and the second fragment (Fig. 2, right) is unmistak- Graz-Köln, 1966, 13; IVO BOJANOVSKI, Mogorjelo – Rimsko ably the 1ère couronne of a Corinthian capital. Although there Turres, Glasnik Zemaljskog muzeja Bosne i Hercegovine u Saraje- are numerous ways to shape a 1ère couronne in a Corinthian vu, n.s., vol. XXIV, (1969), 137-163; NOËL DUVAL, Mogorjelo: capital, this one distinctly reminds of a capital found at the villa fortifiée ou camp militaire?, Starinar, n.s., book XL-XLI, Mediterranean institute near Split, with the obvious provenance 1989-1990, Beograd, (1991), 253-260; MARIN ZANINOVIĆ, from Diocletian’s palace in Split. If we compare the capital from Mogorjelo od vile do kastruma, Godišnjak za balkanološka ispi- Split with that of Mogorjelo, there is typological resemblance, although the execution of the capital from Mogorjelo seems tivanja 32, (2002), 447-456. Besides these, more papers related cruder. The connection between the two is apparent from the to Mogorjelo and ending with a question mark could be added, way the motif in between the protruding and squirming leaves as for example NOËL DUVAL, Existe-t-il un type de palais du (same in both cases) is formed: by three symmetrical and jux- Bas-Empire de la côte dalmate?, Bulletin de Comité International taposed lobes touching its opposites, while at the bottom of the d’Histoire de l’art 1V, (1969), 13. scheme there is a half-leaf. The same pattern is present in Split 4 EJNAR DYGGVE, HERMANN VETTERS (note 3), 14. and in Mogorjelo, and its form is quite distinctive. In Split, the 5 The first communiqué about the finding of the complex in same motif appears in various and very similar forms. So, it is Mogorjelo was published by Franjo Fiala in 1893. See FRA- not out of the question that the model for this 1ère couronne NJO FIALA, Prilozi rimskoj arhologiji Hercegovine, Glasnik from Mogorjelo was, in fact, a type of capital from Diocletian’s Zemaljskog muzeja u Bosni i Hercegovini V, 3, (1893), 726-527; palace in Spalato. See DANIELA MATETIĆ-POLJAK, Les chap- KARLO PATSCH, Zbirka rimskih i grčkih starina u bos.-herc. iteaux du palais de Dioclétien, Dioklecijan, tetrarhija i Dioklecija- Zemaljskom muzeju (sa 147 slika u tekstu i 2 table), Glasnik nova palača. O obljetnici 1700 godina postojanja, Zbornik radova Zemaljskog muzeja u Sarajevu XXVI, (1914), 141-220. The part s međunarodnog simpozija održanog od 18. do 22. rujna 2005. concerning Mogorjelo covers pages 157 to 161. u Splitu (eds. N. Cambi, J. Belamarić, T. Marasović), Književni 6 Such as Karl Maria Swoboda, Franz Oelman, or Eduard Anthes. krug, Split, (2009), 215-216, 218, Fig. 50, 55, 56. As to the other They created and propagated an image of a fortified villa on the fragment, it is only partly preserved and documented only by a basis of comparisons with residences from the northern prov- rather crude photography. Still, it is obvious that the fragment inces. See KARL MARIA SWOBODA, Römische und romanische belonged to a Corinthian capital with an extremely protruding Paläste: eine architekturgeschichtliche Untersuchung, Kunstverlag angular volute, beneath which the remains of a thick protruding Anton Schroll, Wien, 1919. However, the most problematic is leaf are visible in the photograph. It is difficult to make precise Dyggve’s reference to Josef Strzygowski, whom he called a “giant comparisons, but a similar combination of a protruding volute who had risen above the standing water of art history,” as Strzy- and a protruding leaf is present on some of the capitals from gowski was, and became, one of the major advocates of “Gothic the prothyron in Split. See EJNAR DYGGVE, HERMANN VET- theory” and applied it even to regions like Croatia, what natural- TERS (note 3), Tafel VII, 3, 4; ly and quite rightly provoked many, including Ljubo Karaman. 13 EJNAR DYGGVE, HERMANN VETTERS (note 3), 13. See EJNAR DYGGVE, HERMANN VETTERS (note 3), 14. 14 NOËL DUVAL (note 9), 277. 7 He soon became a builder of cottages that organically fitted the 15 In this instance, we thank our reviewer for pointing to us that area around Helsinki. At the International Architecture Exhi- this conclusion of Duval’s is based on information from Patsch’s bition in Ghent in 1921, he won a gold medal for his designs letter to the director of the Museum, Hörmann (from November of that kind. Together with his wife Ingrid Møller Dyggve, he 15, 1899), in which Patsch wrote about the construction tech- also designed furniture. He was a member of the Danish state nique of the oil-press facility (possibly misunderstood by Duval) Naturfredningsrådet (Nature Conservation Council) from 1924 and about “Ein- und Zubauten, die schon durch ihre nachlässige to 1928, and developed a nature conservation plan for the city Fügung und Fundamentlosigkeit sich als ein Product der Deca- of Ghent in 1925. His engagement in archaeology started in dence erweisen.” And although he refers to the stratigraphy af- 1922/23, with his participation in an archaeological expedition terwards, it is certain that the edifice was not built with the same

Nikolina Maraković a Long-Lasting Debate

quality of technique as the fort was. But, Duval’s conclusion has main building to the “piazzetta” in front of it (width 39.6 m) was to be taken with caution. NOËL DUVAL (note 3), 257. also about 1:3. He also noticed that the length of the southwest 16 However, Dyggve’s interpretation of late antique fortification, corridor measured exactly 279 Roman feet (83.7 m). EJNAR like the one in Mogorjelo, as “fortified villae” became a recur- DYGGVE, HERMANN VETTERS (note 3), 19, 30. ring idea in scholarly publications. In the context of Dalmatia, 20 He only documented the postholes, but did not come to the the idea was still alive when Veljko Paškvalin in 2005 inter- conclusion that the beams were carrying the load of the wood- preted the late antique fortress in Karaula (the “Gromile” site) en ceiling. This may be interpreted as cautiousness in making near Kakanj (modern central Bosnia and Herzegovina), as a judgements, but maybe also as intentional by-passing of the fact “fortified villa rustica”. He found a justification for such an in- that the porticoes had wooden ceilings, so thus could have not terpretation of this 3th-4th century fortress in some examples carried the floor above (as Dyggve thought). See EJNAR DYG- from Pannonia and in the work of E.B. Tomas. But, just like in GVE, HERMANN VETTERS (note 3), 30. Mogorjelo, the fortress in Karaula overlooked the fertile plain 21 EJNAR DYGGVE, HERMANN VETTERS (note 3), 53. of Karaula, through which one of the major roads passed, as 22 Most apparent in the last two chapters (VII and VIII) of the Pašalić demonstrated, and which was agriculturally exploited book, written by Vetters. See EJNAR DYGGVE, HERMANN even well before the Roman presence. Thus, we are more in- VETTERS (note 3), 53-63. clined towards the observations of Dorothy J. Thompson, who stressed that castella had become a common feature on imperial 23 Vetters noticed a wide staircase at the south-eastern side, in the estates in North Africa already by the 2nd and 3rd century, pro- side wing of the portico of the main building. He assumed that tecting imperial interests and estates. Considering the number it led to the upper floor, but immediately afterwards he again ac- of forts in the Dalmatian hinterland, and Čremošnik lists 263 knowledged that the portico on the ground floor was supported of them, it could be assumed that fortifications were a common by columns with diameter about 0.3 m. From that he deduced feature of the Dalmatian hinterland. Of course, some of them that the columns were about 3 m high, and the whole portico were restored or rebuilt, or new ones were constructed, as the approximately 4.2 m high (he imagined arches over the columns, archaeological finds attest, precisely at the end of the 3rd and in as in Diocletian’s palace in Split). One thing is certain: 0.3 m the first half of the 4th century, for the same reason – to protect thick columns could not have supported a massive upper floor. the imperial estates, resources (mines), provincial economy, and Still, the staircase had to lead somewhere, so Vetters supported infrastructure such as roads, bridges, etc., just as Čremošnik has the idea about the first floor. EJNAR DYGGVE, HERMANN pointed out and demonstrated on examples. As to the dating of VETTERS (note 3), 53. the fortress in Karaula, Paškvalin’s conclusions remained some- 24 ĐURO BASLER, Arhitektura kasnoantičkog doba u Bosni i Her- what ambiguous, as he at one point dated the transformation of cegovini, Sarajevo, 1972, 38 (Fig. 9). the villa into a fortress some time between the 3rd and 4th cen- 25 See Basler’s footnote 4 on page 36 of his book. ĐURO BASLER turies, and then went on to assume that the fortress could have (note 24), 38. been built at the time of Constantius II. However, the excavated 26 EJNAR DYGGVE, HERMANN VETTERS (note 3), 12. Cy- part of the fortress, with its protruding rectangular tower and presses (Cupressus sempervirens L.), however, do not fit into the indications of inner cubicula, certainly shows features typical of reconstruction, as they were the result of the intervention of the “Diocletianic type” of quadriburgium. About this type, see Karl Patsch in 1903, who had the idea to protect the marshy and below. And that this was indeed primarily a military installation swampy site. On the protection of the site, see ĐURO BASLER, is confirmed by the find of spherical stone ammunition (per- Konzervatorski zahvat na Mogorjelu. Kratak osvrt na problema- haps for the organa?), similar to the one found in Mogorjelo. tiku radova (Traveaux de conservation à Mogorjelo), Naše star- See VELJKO PAŠKVALIN, Zaštitna iskopavanja na antičkom ine I, Sarajevo, (1953), 145-150. lokalitetu “Gromile” u Karauli kod Kaknja, Glasnik Zemaljskog muzeja Bosne i Hercegovine u Sarajevu, 46, (2005), 131-154; 27 He attributed this undertaking of the restoration of oil presses EDIT B. TOMAS, Villa settlements. The Archaeology of Roman etc. to “common Roman economy,” i.e. practicality. ĐURO BA- Pannonia, Akadémiai Kiadó, Budapest, (1980), 275-321; DORO- SLER (note 24), 40. He also published a short review of (Dyg- THY J. THOMPSON, Imperial Estates, The Roman World, Vol. gve’s and) Vetters’ book in 1970, quite illustrative of his conclu- II (ed. J. Wacher), Routledge & Kegan Paul, London-New York, sions about their work and Mogorjelo itself. Some of his claims (1987), 557; IRMA ČREMOŠNIK, Rimska utvrdenja u BiH s should be subjected to criticism. For example, he scolded Vetters osobitim osvrtom na utvrdenja kasne antike, Arheološki vestnik, for being too “rigid” in documenting the archaeological remains 41, Ljubljana (1990), 355-364; ESAD PAŠALIĆ, Antička naselja and praised Dyggve’s reconstruction, which in his opinion was i komunikacije u Bosni i Hercegovini, Zemaljski muzej, Sarajevo, unquestionable because it had been made by an “experienced 1960, 47, 50. architect and archaeologist.” Most of his other remarks are of the same kind, especially those related to the description of the 17 EJNAR DYGGVE, HERMANN VETTERS (note 3), Beil IX. position of the complex. He stressed that there was an earlier 18 As if he had never read Hippolyte Taine, whose thoughts are layer beneath the edifice from the 4th century, and even made fundamental for scholarly research, and especially for historical an imaginative interpretation of the natural environment of the studies. For example, HIPPOLYTE TAINE, Essai sur Tite-Live, complex at that time. He further claimed that the main gate of Librairie Hachette et Cie, Paris, 1896 (especially p. 49). the complex was the western one, implying that Mogorjelo was 19 For example, he acknowledged that the depth of the main build- oriented exclusively to the general direction of Narona. Howev- ing and its width had the ratio of 1:3, and that the ratio of the er, from Fig. 10 it is clearly evident that Narona was not the next

Nikolina Maraković a Long-Lasting Debate

or exclusive destination from Mogorjelo. According to Patsch, Posebna izdanja knj. XCV, Odjeljenje društvenih nauka knj. 27, Sergejevski, and Bojanovski, the network of roads in the Neretva Sarajevo, (1991), 55-65; MARIE-BRIGITTE CARRE, PATRICK delta was quite intricate. It seems that Basler adjusted the facts MONSIEUR, STEFANIA PESAVENTO MATTIOLI, Transport to his preconceptions. See ĐURO BASLER, Osvrti i prikazi: Ej- amphorae Lamboglia 2 and Dressel 6A: Italy and/or Dalmatia? nar Dyggve i Hermann Vetters: Mogorjelo, Glasnik Zemaljskog Some clarifications, Journal of Roman Archaeology 27, (2014), muzeja, n.s. A XXV, Sarajevo, (1970), 213-215. 417-428; GEORGE VELENIS, Συμπραγματευόμενοι Ρωμαῖοι σὲ 28 Duval also criticizes the conclusions made by Ivo Bojanovski af- μία νέα ἐπιγραφὴ τῆς Θεσσαλονίκης, Tekmeria 2, (1996), 8-15. ter the restoration campaign in this quadrant in 1963-1965. As 32 IVO BOJANOVSKI (note 3), 137-163. To complement his con- he points out (repeating the conclusions of KARLO PATSCH, clusions, we might add that such fortified structures, according note 5), the oil press belongs to a much later period, close to to Bajenaru’s classification, would fall into the following catego- the building of Christian churches. His opinion that the oil ry: “Small castellum-type (praesidium / centenarium / castra / press and other installations for production are much later than castellum / phrourion)… during the Late Empire, though there the castellum seems right, and we agree that such installations are different shapes of fortifications, rectangular or polygonal, would not be expected in a castellum from the late 3rd or early that usually serve a military purpose. We consider that in this 4th century. See NOËL DUVAL (note 3), 257. See also IVO BO- category should be included fortifications with a surface of more JANOVSKI, Antička uljara na Mogorjelu i rekonstrukcija njenog than 0.40 ha and up to 1 ha, which constitutes the upper limit of tlocrta (L’huilerie antique de Mogorjelo et la reconstruction de minor fortifications.” The size of Mogorjelo was 0.977 ha, which son pressoir), Naše starine XII, (1969), 27-53. conforms to this category. Considering the form of the complex, 29 ĐURO BASLER (note 24), 42. Bojanovski’s comparisons were correct and could be further broadened by examining Bajenau’s catalogue. See CONSTAN- 30 When it comes to the Roman forts, fortlets, fortified manors, TIN BAJENARU (note 30), 53, 220. etc., the terminology is quite complex, and it becomes especially complex with the age of the Tetrarchy. See CONSTANTIN BA- 33 ĐURO BASLER, Mogorjelo, Arhitektura, Urbanizam 31 [časopis JENARU, Minor Fortifications in the Balkan-Danubian Area from za arhitekturu, urbanizam, primenjenu umetnost i industrijsko Diocletian to Justinian, Editura Mega, Cluj-Napoca, 2010, 52. oblikovanje], 6, (1965), 10-11; IVO BOJANOVSKI (note 3), 161. 31 This shall be an important point in this paper. We certainly agree 34 IVO BOJANOVSKI (note 3), 161. that Mogorjelo served a similar purpose, but not exactly the one 35 Here we have to add the conclusions of Irma Čremošnik about that Basler imagined. As we will argue, the castellum may have the ceramic finds from Mogorjelo. See IRMA ČREMOŠNIK, been in a moist environment at the end of the 19th century, but Keramika iz rimskog nalazišta Mogorjela, Glasnik Zemaljskog there are a lot of arguments that the surroundings of the castel- muzeja u Sarajevu, n.s., VII, Sarajevo (1952), 241-271. lum were quite different in ancient and late antique ages. Thus, 36 IVO BOJANOVSKI, Bosna i Hercegovina u antičko doba, Cen- considering Busuladžić’s systematisation of agricultural tools tar za balkanološka ispitivanja, book 6, Sarajevo, 1988, 125. It (see note 43) found on the site, as well as other finds from the is strange how Bojanovski used the terms villa rustica and pa- area, it becomes apparent that the area around the fort was more latium, putting them in the same line. In fact, it seems that he than probably cultivated and agriculturally exploited through- had taken over Dyggve’s position once again. It is highly prob- out the antiquity and late antiquity. Existence of an olive press, able that he did not think over these two terms, otherwise he constructed either before of after the construction of the castel- would have noticed that they were in direct opposition to his lum (see note 15), does not change the fact that the area was statement from 1969, that Mogorjelo was not a praetorium fundi. intensively agriculturally exploited for the cultivation of olives, See above. and probably wine, at least from the second half of the 1st cen- tury BC. After all, Mogorjelo is in direct vicinity of Čapljina, i.e. 37 NOËL DUVAL (note 3). Tasovići, where at the second half of the 1st c. BC brothers Gaius 38 NOËL DUVAL (note 3), 259. Papius Celsus and Marcus Papius Kanus (CIL III, 14625) obvi- 39 Duval explained everything he had against Dyygve’s approach ously established a profitable business of wine production. As in 2003. NOËL DUVAL (note 9). Dyggve really was somewhat Nenad Cambi assumed, they had even had their own production opinionated and superficial in many respects. If we look back at of amphorae of the Lamboglia 2 type, although this conclusion his theories about Marusinac, the episcopal complex in Salona, was contested in 2014 by Marie-Brigitte Carre, Patrick Monsieur, his interpretation of the first oratoria in Salona, his interpreta- and Stefania Pesavento Mattioli. Leaving the amphorae aside, it tion and reconstruction of the building complex in Polače, or is certain that the Papius brothers had a market-orientated es- his preconceptions about the peristyle in Diocletian’s palace in tate of some extent in the area already by 36 BC. We might only Split, etc. we will conclude that most of his theories and inter- wonder if they belonged to the same wealthy merchant family pretations missed the point. Considering that Dyggve’s “Gothic” of Papii that established itself so well in Thessaloniki, almost at theories are still taken seriously by some scholars, Duval’s some- the same time. About the hydrography of the Naronitan area, what harsh critique seems justified. see also BARTUL ŠILJEG, Neki antički lokaliteti naronitanskog 40 NOËL DUVAL (note 3), 255. agera, Izdanje Hrvatskog arheološkog društva 22, (2003), 267- 276; JAKOV VUČIĆ, Topografija naronitanske biskupije, PhD 41 MARIN ZANINOVIĆ (note 3), 447-456. dissertation, University of Zagreb, 2012, 15-19, Fig. 2. See also 42 Similar terminological confusion, which stems from the general NENAD CAMBI, Amfore kasnorepublikanskog doba i njihova confusion about Mogorjelo, is reflected in the otherwise excel- produkcija u Dalmaciji, Akademija nauka i umjetnosti BiH, lent work on villae in the territory of Bosnia and Herzegovina by

Nikolina Maraković a Long-Lasting Debate

Adnan Busuladžić. He did notice that the fort, indeed, should be lers and boars teeth, which would suggest that even the marshy regarded as a castellum (in the sense of a fortified villa for him), parts were used as hunting grounds. And it should be noted, as and he even made a well justified comparison with the partly our reviewer suggested, that there were possibly canabae in the excavated remains of the fort in Karaula near Kakanj; he then vicinity of the fort, as suggested by Patsch in his letters and by reverted to calling it a castrum, or much more frequently a villa Bojanovski in his review of Dyggve’s and Vetters’ book. See IVO (of a developed ground plan). In all, it appears that he shared the BOJANOVSKI (note 28). age-old dilemma about the nature and function of Mogorjelo. 48 Except for Basler’s publication, the majority of them are printed See also note 15 in ADNAN BUSULADŽIĆ, Rimske vile u Bosni with incorrect orientation. i Hercegovini / Roman villas in Bosnia and Herzegovina, Zem- 49 Not on the western side, as Basler presumed. See note 27. aljski muzej Bosne i Hercegovine, Sarajevo, 2011, 70-71. 50 Of course, both ortostates were obviously reused fragments of some 43 EDIN VELETOVAC (note 1), 329. older building, appropriately reused to symbolically accentuate the 44 Adnan Busuladžić cautiously concluded that the finds of stone nature of the new edifice. See ĐURO BASLER (note 24), 40. projectiles in Mogorjelo pointed to the installation of an ona- 51 Or they were present in limited numbers, as Patsch and Bo- gra, although of the kind that used smaller calibre projectiles. janovski mentioned canabae. See note 47. A wooden construction of that piece of military equipment was not found, but per analogiam with other, similar finds, it 52 IVO BOJANOVSKI (note 3). could be deduced that such a weapon existed in Mogorjelo. As 53 ADNAN BUSULADŽIĆ (note 44). the onagra came into standard military use in the 4th century, 54 Lukas de Blois has vividly demonstrated the effects of the 3rd- as the only stone-throwing weapon, this only corroborates the century crisis, as it affected even the military ranks, which suc- conclusion that Mogorjelo was indeed a complex with military cumbed to banditry along the main military transit routes. De presence. Busuladžić was also inclined to date some other piec- Blois reminded of the complaints about military avarice and es of military equipment to the period after the erection of the misbehaviour, which brought villagers from Takina, Skaptopare fortress, and his conclusion would conform to the general line and Aragoe, places in the Balkans and Asia Minor, into misery of development of the site. See ADNAN BUSULADŽIĆ, Neki and bankruptcy during the reign of Caracalla, Gordian III, and primjeri rimske vojne opreme iz Mogorjela, Radovi kolokvija Philip the Arabian well before the intense military activity in the Rimska vojska u procesu romaniziranja provincije Dalmacije period between 249 and 284. So, looting and pillaging was that Sinj, 13. listopada 2006., Izdanja Hrvatskog arheološkog društva, internal threat with which the emperors had to deal at the end vol. 27 (eds. Anita Librenjak, Domagoj Tončinić), Zagreb, 2011, of the 3rd century, when the crisis became even more acute after 345-361. His analysis of fibulae from Mogorjelo resulted, quite the fall of Dacia. Thus, it would be expected that Diocletian, one correctly, with the same conclusions. A number of crossbow of the best organisers of all times, intervened and established brooches, as well as ring fibulae and penannular brooches, all more secure routes across the Empire. See LUKAS DE BLOIS, from the second half of the 3rd or the first half of the 4th century, The Crisis of the Third Century A.D. in the Roman Empire: A indicate military presence. As Busuladžić stressed, the type of Modern Myth?, The Transformation of Economic Life under the crossbow brooches is associated with military men; so according Roman Empire. Proceedings of the Second Workshop of the Inter- to him, they are additional evidence that Mogorjelo functioned national Network Impact of Empire (Roman Empire, c. 200 B.C. – as a military installation between the 4th and 6th centuries. See A.D. 476), Nottingham, July 4-7, 2001, (eds. L. de Blois, J. Rich), ADNAN BUSULADŽIĆ, Zbirka fibula iz Mogorjela / The Fibu- Brill, (2002), 204-217. lae Collection from Mogorjelo, Opuscula Archaeologica 32, Za- 55 See, for example, JAROSLAW BODZEK et al., Results of “Ar- greb, 2008, 21-54. chaeological Study of Dajaniya & Tuwaneh” (ArTu:DTu) 2018 45 This was overlooked by the majority of scholars. The major pro- survey of Dajaniya (Ma’an-Husseiniyeh), Southern Jordan, Dis- jects of draining the Neretva delta were started in 1881 and last- covering Edom: Polish Archaeological Activity in Southern Jordan, ed until 1889. But, systematic and more extensive melioration of (ed. Piotr Kołodziejczyk), Krakow, (2018), 51-68. the valley was conducted from the 1950s to the 1980s, when the 56 MERSIHA IMAMOVIĆ, BEGO OMERČEVIĆ, Rimska putna area finally became fertile. See NIKŠA BOŽIĆ, MARIJANA ZL- komunikacija Salona-Argentaria, Historijska misao, IV, Tuzla, ODRE, Revitalizacija ruralnih naselja doline Neretve. Smjernice (2018), 14; ESAD PAŠALIĆ, Period rimske vladavine do kraja za integralnu zaštitu ruralnih krajolika i održivi razvoj turizma III vijeka naše ere, Kulturna istorija Bosne i Hercegovine, Sara- delte rijeke Neretve, Regionalna razvojna agencija Dubrovačko- jevo, (1984), 191-236. neretvanske županije – DUNEA, Dubrovnik, 2019. 57 It should be taken into consideration that transport along mili- 46 In fact, it seems that Patsch was the only one who was aware of tary routes exerted great pressure on the local population; either the quality of the surrounding terrain. because the provisions for the entourage were requisitioned from 47 ADNAN BUSULADŽIĆ, Antički željezni alat i oprema sa pros- them (oxen or other beasts of burden), or simply because of tora Bosne i hercegovine / Iron Tools and Implements of the Ro- plunder. Complaints of these communities against unauthorized man Period in Bosnia and Herzegovina, Univerzitet u Sarajevu, military seizure of men, animals, waggons, and hospitality in bil- Zemaljski muzej Bosne i Hercegovine, 2015, 14-15, with a cata- lets and other facilities for state transport are a recurrent theme logue. Further still, we thank our reviewer for the suggestion in Roman history, as Stephen Mitchell has pointed out. That is that the swamps were not a completely unexploited resource, as why imperial decrees regulating such affairs were issued so of- Patsch noted in his letter about the frequent finds of deer ant- ten. Furthermore, besides the security of transport, one should

Nikolina Maraković a Long-Lasting Debate

mention the fact that the cost of transportation was, in the Ro- etva used to be called T(e)rsana (from the word “arsenal”), the man world, as Paul Erdkamp has stressed, a part of the produc- remains of which have mostly been washed away by the Ner- tion costs. Thus, a longer route would not have been preferable etva or destroyed during the construction of the railroad bridge either from the standpoint of security, or from that of economy. across the river. T(e)rsana was located just to the southwest of About the Pisidian monument and the behaviour of the military the complex in Višići. The sheer toponym would imply the ex- on transport routes, and on the Roman transport in general, see istence of a naval yard. As Patsch noted, its remains had been STEPHEN MITCHELL, Requisitioned Transport in the Roman buried before they were finally destroyed by the building of the Empire: A New Inscription from Pisidia, The Journal of Roman bridge, which would indicate that they belonged to an era long Studies Vol. 66 (1976), 106-131; ANNE KOLB, Transport und before the Middle Ages. The location is, indeed, quite conveni- Nachrichtentransfer im römischen Reich, Klio: Beiträge zur alten ent for a naval yard. If there was a naval yard (arsenal) adja- Geschichte, Beiheft, n.F., Bd. 2. Berlin, 2000, 54-63, 71-82; About cent to the complex in Višići, that would explain a lot about the general rules of using diplomata on Roman roads, see LUKAS the disposition (baths, hospitia, shops, etc.) and development of LEMCKE, Status Identification on the Road: Requisitioning of the complex. Along with this, one should bear in mind that the Travel Resources by Senators, Equestrians, and Centurions with- medieval local Bosnian parish was called Luka (“port”) and that out diplomata. A Note on the Sagalassus Inscription (SEG XXVI it included both Višići and T(e)rsana. In addition, one should 1392), GEPHYRA 9, (2012), 128-142. See also PAUL ERDKAMP, be reminded that the Bosnian king Tvrtko I built a shipyard The Grain Market in the Roman Empire. A Social, Political and and a port in the relatively nearby Opuzen, obviously respect- Economic Study, Cambridge University Press, 2005, 112-113. ing the logic of sailing in the Neretva delta. One of the places that belonged to the parish of Opuzen was called Lađište (lit- 58 The fact that the Roman route leading from central Bosnia, erally “shipyard”). Folk stories preserve the memory that this across the modern town of Mostar, to the Neretva delta gained was the place where large vessels use to anchor in the “olden in importance as the shortest route was also noticed by Ivo Bo- days” and that this was the actual harbour of Narona. But, the janovski. Along the route, a significant number of new milestones main harbour during the Middle Ages was certainly the “Dri- were erected during the 3rd and 4th centuries. As Bojanovski has jeva square” in the town of Gabela, known as the “mercatum noted, it does not mean that the route was built at that time, but Narenti, forum Narneti” and first identified as such by K. Jiriček only that it gained in importance at that time and was probably back in 1879. See extensively about the issue in ĐURO TOŠIĆ, repaired and enlarged. He explicitly writes: “The installation of Trg Drijeva u srednjem vijeku, Biblioteka “Kulturno nasljeđe”, milestones in honour of emperors does not only demonstrate Veselin Masleša, Sarajevo, 1987; so, there are several possible subservient loyalty to the ruler, but also intense traffic, especially locations of the shipyard from which the goods from Mogor- in the 3rd and 4th centuries, when mining improved the general jelo could have been shipped. In any case, the historical con- condition of our province.” Quite correctly, he associated that tinuity of the use of certain appropriate spots should not be undertaking with the intensification of transport of valuable ignored. Finally, it should be stressed that the town of Narona ores, gold and silver, by that route. That route of prime impor- is the least likely candidate. KARL PATSCH (note 55), 387- tance, as it seems, ended actually just north of Mogorjelo, at the 388; MARIJAN SIVRIĆ, Srednjovjekovna župa Luka, Povijest crossroads settlement called Tasovići. The next stop, and prob- hrvatskog Počitelja, (ed. Krunoslav Kordić), Općinsko pogla- ably the last one before the shipment, was Mogorjelo. See KARL varstvo Čapljina, Čapljina-Zagreb (1996), 168-224; DOMAGOJ PATSCH, Pseudo-Skylaxovo jezero. Prinos povijesti donjeg VIDOVIĆ, Pogled u toponimiju župe Opuzen, 300-ta obljetnica poriječja Neretve, Glasnik Zemaljskog muzeja u Bosni i Hercego- Župe svetoga Stjepana Prvomučenika, (eds. Zoran Curić et al.), vini XVIII, book 4, (1906), 367-390; IVO BOJANOVSKI, Prilozi Župa svetog Stjepana Prvomučenika Opuzen, (2016), 168. za topografiju rimskih i predrimskih komunikacija i naselja u rimskoj provinciji Dalmaciji, Godišnjak Centra za balkanološka 63 VACLAV RADIMSKY, Prekopavanje u Domaviji kod Srebren- ispitivanja 17, (1978), 64. ice, godine 1891, Glasnik Zemaljskog muzeja u Bosni i Hecego- vini, IV, book 1, Sarajevo, (1892), 4. 59 See the daily Dnevni Avaz, issue of August 27, 2020. 64 The majority of Villae mit Eckrisaliten (like Mehring, Berlingen- 60 As part of the same project, the old railroad connecting Sarajevo Neuhaus, Bollendorf, Wachenheim, Weinbergshof, Tittmoning with Ploče will be restored. In its final stages, this railroad fol- and most of the other places) do have side risalites and porticos, lows exactly the ancient route through Herzegovina (Sarajevo- but the porticus is mostly elevated and serves as an entrance to Nevesinjsko polje-Mostar-Čapljina/Tasovići-Mogorjelo-Višići- the rest of the building, which is usually arranged around an in- (Ploče). See the daily Večernji list, issue of January 16, 2022. ner courtyard. This was certainly not the case in Mogorjelo. The 61 The dilemma where exactly the harbour of ancient Narona was lo- only similarity between the main building in Mogorjelo and the cated is discussed in the next note. There are multiple indications northern Villae mit Eckrisaliten is the general impression of the that it was situated in or around modern Gabela. See below. façade of the building, but even there the differences are appar- 62 At first, it could be assumed that the ores or any other kind of ent – there is no central staircase in front of the main building, valuable goods were shipped from the port of Narona. However, no elevated portico, etc. although that road was part of the main communication in the 65 MICHAL GAWLIKOWSKI, Les principia de Dioclétien à Pal- area, that would be highly illogical considering the lengthy land myre. Projet et réalisation, Le dessin d’architecture dans les so- route from Mogorjelo to Narona, winding around the northern ciétés antiques, Actes du Colloque de Strasbourg 26-28 janvier edges of the Naronitan marshes. On the other hand, we ought 1984 (Travaux du Centre de Recherche sur le Proche Orient et la to bear in mind that the area between Višići and the river Ner- Grèce Antiques 8), Strasbourg, (1985), 283-290.

Nikolina Maraković a Long-Lasting Debate

66 MICHAL GAWLIKOWSKI, Le camp de Dioclétien. Bilan s., Mélanges de l’École Française de Rome 56, (1939), 178-248; préliminaire, Palmyre, bilan et perspectives, Colloque de Stras- MICHEL REDDÉ (note 67), 101. bourg (18-20 octobre 1973) (Travaux du Centre de Recherche sur 79 This last feature is present in all of the three mentioned examples. le Proche Orient et la Grèce Antique 3), Strasbourg, (1976), 153- 80 Although, we are thankful to our reviewer that he conveyed to 163 (Fig. I). us the information that in the abovementioned letter from No- 67 JAMES LANDER, Typology and Late Roman Fortifications: The vember 14, 1899, Patsch mentioned to the director of the Mu- Case of Diocletianic Type, in Limes 12 Stirling, III, (1980), 1051- seum in Sarajevo that he had found an extensive amounts of 1060; JAMES LANDER, Roman Stone Fortifications: Variation shells (Austerschalen – oyster shells) on the site of Mogorjelo. and Change from the First Century A.D. to the Fourth, BAR Int. That would certainly conform to our assumption that the larger Ser. 206, Oxford, 1984, 181-193; MICHEL REDDÉ, Dioclétien et hall of the main building was indeed a reception room. les fortifications militaires de l’Antiquité tardive. Quelques consi- 81 Numismatic finds could be considered as an indication that the dérations de méthode, Aniquité Tardive 3 (1995), 91-124. estate had become more market-orientated and less orientated 68 JAMES LANDER (note 67, 1984), 181. on small-scale production for the needs of residents and their 69 MICHEL REDDÉ (note 67), 91-124. labour. Since the earliest numismatic finds stem from the age of 70 Which is different from both the so-called “pre-Diocletianic” Trajan and subsequently multiply through the reigns of Hadrian type and from the “post-Diocletianic” one. and Antoninus Pius, they confirm that the owner (conductor or tenant) started to participate in monetary transactions and cash 71 JAMES LANDER (note 67, 1984), Figs. 182, 185, 173, 174, 177. payments, and that certainly means that production on the es- 72 In this case, as in Mogorjelo, we could assume that there were tate had stepped up for a notch. If this was indeed so, then the “northern” influences in the design of the main building. See improvement of production on the estate implied a couple of CONSTANTIN BAJENARU (note 30), Fig. 148. things. First, that fresh capital was invested in the exploitation of 73 CONSTANTIN BAJENARU (note 30), 72. the estate. Secondly, that the economy of the estate had become 74 CONSTANTIN BAJENARU (note 30), 262, Pl. 31. diversified. As Paul Erdkamp wrote: “… wealthy landowners who had money to spare rather bought additional land or they invest- 75 Dated by an inscription. ISRAEL ROLL, A Latin Imperial In- ed it in the other sectors of the villa economy: primarily olive scription from the Time of Diocletian found at Yotvata, Israel oil and wine, but also poultry and fish, and factories for pottery, Exploration Journal 39, (1989), 239-260. Cf. ZEEV MESCHEL, brick or tiles.” We cannot know in how many ways the economy A Fort at Yotvata from the Time of Diocletian, Israel Exploration of the estate in Višići was diversified to maximise exploitation, Journal 39, (1989), 228-238; UZI AVNER et al., The Roman Fort but the addition of a facility for the production of bricks, tiles, at Yotvata: Interim Report (2003), Journal of Roman Archaeology and possibly pottery was part of that process of diversification 17, (2004), 405-412. and orientation towards market. After all, it seems that over time 76 MORDECHAI GICHON, Excavations at Mezad Tamar-Tamara, even the oldest part of the complex (building B) became a place Saalburg Jahrbuch 33, (1976), 80-94; MORDECHAI GICHON, of negotiatio and was re-dressed to impress the negotiatores. In Mezad Tamar / “Tamara”. Vorbericht der Grabungen 1973-1974, consequence, numismatic finds seem as evidence that something Studien zu den Militärgrenzen Roms II. Vorträge des 10. Interna- in the economy of the estate has changed in comparison to the tionalen Limeskongresses in der Germania Inferior, Köln-Bonn, 1st century. IRMA ČREMOŠNIK, Rimska vila u Višićima – Die (1977), 445-452; MORDECHAI GICHON, Developments in römische Villa in Višići, Glasnik Zemaljskog muzeja u Sarajevu, the Research on the Limes Palaestinae during the Last Two vol. XX, (1965), 147-260; PAUL ERDKAMP (note 57), 173. Decades, Limes 17 Zalău, (1999), 243; SHELAGH GREGORY, 82 It should be pointed out that the architecture and fragmentary Was There an Eastern Origin for the Design of Late Roman architectural decoration found in the building D offer just lim- Fortifications? Some Problems for Research on Forts of Rome’s ited possibilities for determining the date of the construction of Eastern Frontier, David Kennedy (ed.), The Roman Army in the this part of the complex. Čremošnik was inclined to date it to East. Journal of Roman Archaeology Supplementary Series, Ann the mid-4th century. However, her conclusion is based on limited Arbor, 1996, 169-209. argumentation. Of course, some of the fragments of architectur- 77 MICHEL REDDÉ (note 67), 100-101, Fig. 19; LOUIS LESCHI, al decoration from the building D are so generic that they can Centenarium quod Aqua Viva appelatur..., Comptes rendus des neither corroborate nor refute the dating to the mid-4th century. séances de l’Académie des Inscriptions et Belles-Lettres 85, (1941), However, three capitals from the room 11D did catch the atten- 2, 163-176; LOUIS LESCHI, Le “Centenarium” d’Aqua Viva près tion of Čremošnik, and those are capitals of the pilasters with de M’doukal (Commune mixte de Barika), Revue Afrique 87, acanthus leaves (T. III. 4). Čremošnik did recognise that such (1943), 5-22. capitals could be found in Salona and Spalato, but in her last 78 As to the date of construction of these forts, Lander is quite sure verdict of their provenance she reached for an, in our opinion, about the one in Seba Mgta, dating it to the Tetrarchic period. erroneous and farfetched conclusion of Edmund Wiegand about There has been some discussion and diverging opinions about the existence of the “western type of acanthus capitals.” But, even Bourada. But, considering the resemblance to other such forts, if there were such a type, still the capitals from Višići would not and the lack of distinctive Constantinian towers, it is more likely conform to it, as the lobes of the leaves touch each other on the that it was a part of the Tetrarchic building endeavour in North capitals from Višići. And if the so-called “western type” is de- Africa. See JAMES LANDER (note 67, 1984), 208; JULIEN fined by rounding or softening the edges of the lobes, then there GUEY, Note sur le limes romain de Numidie et le Sahara au IVe are examples of the same features on capitals from Diocletian’s

Nikolina Maraković a Long-Lasting Debate

palace (see Figs. 63 and 65 in DANIELA MATETIĆ-POLJAK / that the complex had an earlier phase are compelling, especially note 12/). The same could be said about another remark, namely when compared with the development of the complex in Gamzi- that the lobes of the acanthus leaves of the “western type” ex- grad. However, whether it had the proposed function of the gy- tend right down to the bottom of the whole leaf. Again, the same naeceum is a topic open for discussion. It certainly was an impe- mentioned examples from Spalato, as well as many others from rial installation, and it certainly was rather hastily transformed the same palace, diminish the possibility of existence of the so- into the imperial residence, just like Goran Nikšić explained. called “western type of the acanthus capital.” Today, the array However, in view of the relatively recent discoveries of a macel- of possibilities for comparisons has significantly improved, and lum (2013) and a harbour equipped with a dry dock and a pro- thus the basis for attributing the capitals from Višići to some truding platform in front of the “palace” (2006/2007), it seems type of capitals of dubious existence is quite weak. Thus, it seems that the complex had more functions than it had been supposed. that Čremošnik’s first impressions about these capitals were The fact that there was a macellum north of the complex suggests more correct than her final conclusions. As a matter of fact, cer- that the place was much more developed than previously as- tain features make them more like the so-called “eastern type” sumed. As Armando Cristilli has eloquently explained, the crite- – for example, that the lobes of leaves touch those of the sur- ria for building macella had changed by the late antique period. rounding leaves. Of course, the execution of these capitals is far To quote Cristilli: “… the macellum had to be built only in ur- cruder and rustic than the ones from Mogorjelo, not to mention banized sites with good economic development, with a location Spalato. But, in general, they belong to the same group. Finally, capable of connecting with the outside world, with a fair number this certainly does not mean that they were not carved in the of inhabitants and, finally, with a high civic rank.” Thus, it can be mid-4th century, but it is equally likely that they were made well concluded that the complex and its surroundings had all of these before that time. See IRMA ČREMOŠNIK (note 81), 162-163; properties. When we take into consideration the analysis of ce- EDMUND WIEGAND, Die Stellung Dalmatiens in der römis- ramic finds from the “palace” conducted by Ivančica Dvoržak chen Reichskunst, Strena Buliciana, Zagreb-Split, 1924, 91. Schrunk, which shows that the complex was from the beginning 83 See IRMA ČREMOŠNIK (note 81), 158-166. inhabited by residents of higher social status, it is not out of question that the complex had an administrative function before 84 The change in the management of the estate was reflected in it was turned into the Emperor’s retirement palace. If that was the production on the estate. The initial residence now acquired indeed the case, the parallels between Mogorjelo and Split are a new part – a production facility with a furnace (building A) justified. See GORAN NIKŠIĆ, Diocletian’s Palace – design and to the north. The rectangular shape of the furnace led Irma construction, Bruckneudorf und Gamzigrad. Spätantike Paläste Čremošnik to date its construction in the 2nd or 3rd century. und Großvillen im Donau-Balkan-Raum. Akten des Internation- IRMA ČREMOŠNIK (note 82), 152-153. alen Kolloquiums in Bruckneudorf vom 15. bis 18. Oktober 2008, 85 IRMA ČREMOŠNIK (note 82), 199-200. (eds. Gerda v. Bülow, Heinrich Zabehlicky), Bonn, (2011), 187- 86 See IRMA ČREMOŠNIK, Mozaici i zidno slikarsktvo rimskog 202; JOŠKO BELAMARIĆ, GYNAECEUM IOVENSE DALMA- doba u Bosni i Hercegovini, Veselin Masleša, Sarajevo, 1984, TIAE - ASPALATHO, Prilozi povijesti umjetnosti u Dalmaciji, 133-140, 168-178. See also ADNAN BUSULADŽIĆ, Umjetnost Vol. 40, No. 1, (2005), 5-33; JOŠKO BELAMARIĆ, Diocletian’s antičkih mozaika na tlu Bosne i Hercegovine / Ancient Mosaics Palace, PhD dissertation, University of Zagreb, 2009; KATJA on the Territory of Bosnia and Herzegovina, Zemaljski muzej u MARASOVIĆ, Luka Dioklecijanove palače, Pomorski Split do Sarajevu, 2008, 31-38, with a catalogue. početka XX. stoljeća – Zbornik radova, Biblioteka knjiga Med- 87 ELIO LO CASCIO, The Imperial Property and its Development, iterana 108, (ed. Nenad Cambi), Književni krug, Split, (2019), Ownership and Exploitation of Land and Natural Resources in the 77-96; ARMANDO CRISTILLI, Macellum and Imperium. The Roman World (eds. Paul Erdkamp, Koenraad Verboven, Arjan Relationship between the Roman State and the Market-Building Zuiderhoek), Oxford University Press, (2015), 62-70. The change Construction, Analysis Archaeologica. An International Journal in ownership may also have happened later, in Diocletian’s time. of Western Mediterranean Archaeology, vol. 1, Edizioni Qua- His reforms, when contrasted with the economic mores of the sar, Roma, (2015), 69-86; IVANČICA DVORŽAK SCHRUNK, earlier ages, are best described in J.W. Ermatinger’s dissertation: Dioklecijanova palača od 4. do 7. stoljeća u svijetlu keramičkih “The two most important were the change in land holding pat- nalaza, Vjesnik Arheološkog muzeja u Zagrebu, 3rd s., XXII, terns and the continuing involvement of the state in the lives of (1989), 91-105. the private citizen. This is best seen in the papyri from Panopolis 89 See VANJA KOVAČIĆ, Kasnoantička vila s portikom na Mirju from 298 and 300 A. D. Archaeological evidence serves as another kod Postira, Klesarstvo i graditeljstvo, Vol. XXI, No. 3-4, (2010), major source of information for the ancient historian. The exten- 25-37; TIN TURKOVIĆ, The Late Antique “Palace” in Polače sive building activities all over the empire, dating from Diocle- Bay (Mljet) – Tetrarchic “Palace”?, Hortus Artium Medievalium tian’s reign, point to a revival in municipal rebirth, in contrast to 17, (2011), 211-233. the earlier period.” Indeed, Ermatinger’s dissertation thoroughly examines and describes the pattern of Diocletian’s reforms. See JAMES WILLIAM ERMATINGER, The Economic Reforms of Dio- cletian, PhD dissertation, Indiana University, 1988, 19. 88 We have to agree with Joško Belamarić and Goran Nikšić in that the so-called Diocletian’s palace had a phase before it was turned into the Emperor’s retirement residence. Their arguments

Most read articles by the same author(s)

Similar Articles

1-10 of 57

You may also start an advanced similarity search for this article.