Roman Rural Architecture in the Territory of Colonia Iulia Pola

Roman Rural Architecture in the Territory of Colonia Iulia PolaAuthor(s): Robert MatijašićReviewed work(s):Source: American Journal of Archaeology, Vol. 86, No. 1 (Jan., 1982), pp. 53-64Published by: Archaeological Institute of AmericaStable URL: http://www.jstor.org/stable/504293 .Accessed: 06/11/2011 12:58

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

Archaeological Institute of America is collaborating with JSTOR to digitize, preserve and extend access toAmerican Journal of Archaeology.

http://www.jstor.org

Roman Rural Architecture in the Territory of Colonia lulia Pola

ROBERT MATIJASIC

Abstract General problems concerning Roman villas and the Roman economy in Histria (Yugoslavia) are followed by a discussion of nine Histrian villas, mainly exca- vated before 1914. Three types of villas-luxurious, farm-type, and mixed-are distinguished. The farm- type, with a manufacturing quarter, can have two sub- types, i.e. with specialized or mixed processing facili- ties. The third type has a mixture of agricultural func- tion and luxurious appointments.

The Roman rural villa developed as a contrast to the urban villa, when de-urbanization was taking place in the Roman world. Roman society, although rural in its origins, had previously favored urban

growth as a means to strengthen its political hold on the Mediterranean. The consequence was a rapid expansion of towns in the 1st century B.C. and the establishment of rural settlements in their territo- ries.

Functional classifications of Roman rural villas are based on similarities with urban villas in such elements as architectural decorative schemes.' The three main types of Roman rural villas are:

1. The luxury retreat may have a farming establish- ment incorporated into the architectural layout, but it nevertheless represents a continuation of the urban lifestyle. It also expresses the survival of some of the basic and unchangeable principles of Roman life: individuality and enjoyment of the beauties of nature.2

2. A far more common category is the "urbano-rural" or working villa,3 with a small residential section. Here the pars fructuaria is merged with the pars urbana;4 the majority of villas in the Roman world fall into this group.

3. A third category is what Rostovtzeff called "an agri- cultural factory run by slaves,"s which was the pillar of the Roman economy in the Mediterranean basin, its activities consisting of the intensive processing of agricultural and other products.

This article examines the evidence for the three basic types of Roman rural villa in the Colonia lulia Pola, summarizing archaeological work before 1960. The villas discussed here were excavated without the attention to detail necessary for com-

plete reconstruction and explanation. The remains in situ and the finds were published only superfici- ally; since the former have been covered over and the latter were lost during the two World Wars, the

following summary and interpretation are of neces-

sity incomplete. The attempt to put these poorly known villas of southern Histria into archaeological context has been made to provide some background for the more systematic and comprehensive excava- tions of Roman villas undertaken by Mlakar during the 1960s and 1970s.

At the tip of the Histrian peninsula, the ager of Colonia Pola was not on the main routes leading from Italy to the Balkans and Pannonia. After the establishment of Aquileia in 181 B.C., the Romans subdued the Illyrian Histri in 177 B.C., but the two colonies were not founded until much later: Pola in the time of Julius Caesar, Parentium at the begin- ning of the 1st century A.C.6 In Augustus' division of Italy into ten regions, Histria became part of

Regio X (Venetia et Histria), an action that opened the way to radical Romanization and drew colonists to the fertile ager centuriatus. Colonia Pola reached its apex during the first half of the 1st century A.C., with the zoning of its area.7 The town had two the-

'G.A. Mansuelli, Le ville nel mondo romano (Milan 1958) 18-22; J. Percival, The Roman Villa, An Historical Introduction (London 1976) 54-57, with extensive bibliography. and M. Suit, Anticki grad na istocnom Jadranu (The Classical City on the Eastern Shore of the Adriatic Sea) (Zagreb 1976) 216-19, J. McKay, Houses, Villas and Palaces in the Roman World (Lon- don 1957).

2 Mansuelli (supra n. 1) 16. 3 Percival (supra n. 1) 56. 4 Mansuelli (supra n. 1) 18, and D. Scagliarini, "Ravenna e le

ville romane in Romagna," in Collana di quaderni di antichita

ravennati, cristiane e bizantine (Ravenna 1968) 11. I Percival (supra n. 1) 56; M. Rostovtzeff, Social and Eco-

nomic History of the Roman World2 (London 1957) 552. 6 A. Degrassi, II confine nord-orientale dell'Italia Romana

(Diss. Bern 1954). 7 General works on Histrian antiquities and Roman monu-

ments in Pula are: A. De Ville, Descriptio portus et urbis Polae (Venice 1663); T. Allason, Antiquities of Pola (London 1819); L. Cassas and G. Lavall&e, Voyage pittoresque et historique de l'Istrie et la Dalmatie (Paris 1802); R. Weisshdiupel, Die rom- ischen Altertiimer in Pola (Pola 1892); M. Mirabella Roberti,

54 ROBERT MATIJAICd [AJA 86

aters, an amphitheater accommodating 20,000, a fo- rum with temples and a number of luxurious urban villas with mosaics, the last preserved under the houses of medieval and modern Pula.

In the Roman economy, agricultural investments became increasingly more necessary and more prof- itable, in keeping with the need for a more rational

exploitation of land, and with a growing demand for agricultural products within the state.8 Histria was not well suited to agriculture, but rather to stock breeding. The peninsula was also one of the foremost producers of wine and oil. Conditions for growing grapes and olives are most favorable along the low-lying western coast, on the colonial agri of Pola and Parentium. Histrian products were widely exported, primarily to Pannonia and Noricum, but to the south as well.9 This trade reached its peak in the 1st century A.C., when ceramic workshops were active in Fa'ana (Pansiana) near Pola, and in Cer- var-Loron near Parentium, producing amphorae for oil and wine.10 Pliny the Elder places Histria sec- ond in a list of major oil-producing regions within the Roman world, together with Hispanic Baetica, and second only to the ager Venafrum."1

Numerous remains of olive-oil and wine presses are witness to a sophisticated process for the ex- traction of oil and wine. There is almost no im- portant villa in Histria without its own processing facilities for agricultural products. End-products were intended primarily for trade, by sea and land routes. The distribution of Roman port facilities indicates the importance of the sea routes along the Histrian coast, with Aquileia as the center.12 From this important harbor roads led to Noricum and Pannonia, so that Histrian products, packed in am- phorae, reached the northeastern corner of the Ro- man Empire.

Rural villas in the hinterlands supported this wide-ranging trade by providing the wine and oil. We are fortunate to have the remains of some of these villas, the physical expression of the economic basis of Histria. 1. KOLCI HILL: Registered in the older literature as Monte Collisi, it is situated in the northern part of the island of Veliki Brion, in a slightly elevated position with a panoramic view of the archipelago and the coast. (Italian names are given because the region is still bilingual, and the sites usually appear under them in the archaeological literature.) It was excavated in 1907;13 a three-winged building around the central yard B (ill. 1,B) has an open

o 5 1o

I1

A B

E

O0000

Ill. 1. Kolci Hill, plan of villa

fourth side. Accommodations for workers, slaves and the foreman were in the southwest wing A, while the central part, divided by a passage D, consists of

L'Arena di Pola2 (Quaderni-guida di Pola, Pola 1943); ?. Mla- kar, Die Romer in Istrien (Kulturhistorische Denkmiler in Istri- en 5, Pula 1966); S. Mlakar, Ancient Pula, Cultural and Histor- ical Monuments of Istria 24 (Pula 1974). Most of the recent works have extensive bibliographical references.

8 Mlakar, Die Romer (supra n. 7) 47; see also F. Zevi, "An- fore istriane ad Ostia," Atti e Memorie della societa istriana di

archeologia e storia patria (hereafter cited as Attilstr) n.s. 15 (1967) 31-57.

9 Degrassi, "L'esportazione di olio e olive istriane nell'etai ro- mana," Attilstr n.s. 4 (1956) 104-16.

10 For Monte Loron, see C. Gregorutti, "La figulina imperiale Pansiana di Aquileia e i prodotti fittili dell'Istria," Attilstr 2 (1886) 219-53. Recent work in the field: V. Jurkii, Scavi in una parte della villa rustica romana a Cervera presso Parenzo (I)

(Atti del Centro di ricerche storiche-Rovigno 9, Fiume-Trieste 1978-79) 265-98.

11 Principatum in hoc quoque bono obtinuit Italiae toto orbe, maximae in agro Venafro ... reliquum certamen Histriae terrae et Baeticae par est, Pliny, HN 40.8.

12 A. Degrassi, "I porti romani dell'Istria," in Studi in onore di C. Anti (Florence 1955). Reprinted in Attilstr n.s. 5 (1957) 24-81.

13 A. Gnirs, "Istrische Beispiele fur Formen der antik-r6m- ischen Villa Rustica," Jahrbuch flir Altertiimskunde 2 (1908) 134-35; the identification of various rooms is based on Gnirs' observations. Also see more recent works: ?. Mlakar, Muzejsko konzervatorski zahvati na oto'ju Brioni (Muzeji br. 11-12, Za-

greb 1956-57), and S. Mlakar, Brioni, monografija (Brioni 1971) 30.

1982] ROMAN RURAL ARCHITECTURE 55

a cella olearia with four presses and auxiliary room F (kitchen and praefurnium). A duct leads under the presses from the olive-oil production section of the villa to a smaller basin for sedimentation, and subsequently into the huge storage-room C. The

courtyard is surrounded by a low wall with remains of columns of the peristyle (diameter 0.45 m.).

Finds were rather meager: sherds, and fragments of roof and of floor tegulae of the 1st century A.C. The walls were originally 0.45-0.50 m. wide, and were consolidated at some time after construction so that at some points they are 0.90 m. wide. 2. VERIGE BAY: Known in Italian as Val Catena, it is one of the most important and most monumen- tal sites of the North Adriatic. The ground plan (ill. 2) of the complex is a well balanced arrangement

3,27) leading in from an area with a roof supported by columns. Here was the meeting point of two routes of internal movement: from the temple area and from the cryptoporticus (44) (through passages m and d).'4 From the vestibule (27), a passage (28) led to rooms 29 and 30 with another passage to the eastern manufacturing section of the villa (via stairs in room 17). From room 29, along the thermal section of the villa, stairs (33) led to the residential area. This part of the villa was set around a peri- style (38); the wings of it flank only the southern and part of the western sides, (37,36), with rooms Gnirs calls triclinium (22) and auxiliary chambers (39,40), where fragments of mosaics were found.

The row of rooms extending along the western side was identified by Gnirs as the thermal section. The center is obviously the apsed room (34) in which part of the mosaic floor as well as part of the heating system were recovered.'" The function of the other apse (31) is not clear, but it nevertheless belongs to the "thermal section," especially because it has no apparent communication with the small room immediately adjacent (27: main western en- trance). There are stairs (later closed) leading from room 33 to the rooms around the triclinium (22) and peristyle (38), situated on a higher terrace.

The industrial section of the villa is situated around peristyle 1, on the second terrace,'6 and is connected with the main passage (21) by means of a short and narrow corridor (5). Room 9 could have been, according to Gnirs, a bedroom for slaves or a wine storage-room."7 As it is between the kitchen (6) and the "forum" for crushing grapes, it was more likely an area for collecting grapes before pro- cessing. The rooms in the southeastern corner (10, 11, 12) were used in the production of wine: presses in room 10, sedimentation basin (11), storage room (12) with partially preserved huge vessels. Beneath the peristyle yard (1) there is a cistern for rain water, with the capacity of 231 cubic m.'8

The terrace (23) overlooked the sea, and sup- ported a portico along rooms 24, 25 and 26. Im- mediately above, the second-floor portico articulates rooms 28, 18, 20 and 19. From here, stairs (17 and 33) lead to the third floor. In the eastern part of the villa a wall encloses a huge garden/yard, and the

C -

A

D B

UVALAVERIG E H

0 50 100 150m

Ill. 2. Verige Bay, plan of villa

around a bay on the eastern shore of Veliki Brion. The entire seaside was built up as an embankment, preserved in only a few places; it is mostly under water (H). On slightly downsloping ground build- ings were then erected around the sea wall: a villa with two peristyles (F), a temple area (E), a crypto- porticus (D), a thermal complex with its palaestra (C) and another villa (A). The complex contained every element necessary for sustaining life: presses for the oil and wine, basins for fish (B) and a garden (G). It also included all the features desir- able in a residential setting, such as temples, ther- mae, a palaestra and libraries.

One of the two entrances to the so-called villa with two peristyles was through a vestibule (ill.

14 A. Gnirs, "Forschungen fiber antiken Villenbau in Sud- istrien," OJh 18 (1915) 112.

"5 Gnirs (supra n. 13) 115, 120.

16 Gnirs (supra n. 13) 115. 17 Gnirs (supra n. 13) 136. 18 Gnirs (supra n. 13) 138.

56 ROBERT MATIJASIC [AJA 86

same wall forms a rather narrow passage along the southern side of the complex. On the western end is a huge cistern for collecting water from a spring on the opposite side of the bay (ill. 2, F).19

spectent, ideo quod vespertina lumine opus est uti.20 The residential section might therefore be the "sum- mer residence," owing to its cool position. Light to the inner rooms was provided through openings in

m 44

e d 23

26 25 24

27 28 18 20 19

43

35

S29 17 _F1

1 40

35

16, 42

o 10

o

36 38 1

22 8 7 17 40

rn114 Ill. 3. Verige Bay, "villa with two peristyles," plan

Of the villas known in Histria, the Verige Bay mansion is one of the most luxurious, in spite of the sizeable industrial section. The architecture forms a harmonious belt between the sea and the green un- dergrowth encircling the whole bay (ill. 2). In ex- ploiting the attractive view from the villa to the south shore, the architect decided to ignore the fact that the north side is much colder because of the northern winter wind. The third level is divided into two functional parts: the residential western and the manufacturing eastern sections, although by this arrangement Vitruvius' advice is ignored: Hi- berna triclinia et balnearia ad occidentem hibernum

the peristyles, except for the thermal section which must have had windows in the wall: Omniaque edi- ficia ut luminosa sint, oportet curari. Sed quae sunt ad villas faciliora videntur esse, ideo quod paries nullius vicini potest obstare.21 2a. VERIGE BAY-NORTH: In 1899, during the

building of a road, another villa was discovered im- mediately to the east of the thermal complex of the mansion, (ill. 2, A; ill. 4). According to Gnirs' ob- servations, it seems to have consisted of a cistern and a manufacturing section, while the residential area was spread along the seashore, and was de- stroyed in the subsequent rise of the sea level. The

19 A. Gnirs, "Beispiele der antiken Wasserversorgung aus dem istrische Karstlande," in Strena Buliciana (Zagreb 1924) 129.

20 Vitr., De archit. 6.1. 21 Vitr., De archit. 6.2.

1982] ROMAN RURAL ARCHITECTURE 57

original nucleus was around the cistern (1), near which there is a smaller water reservoir built in an earlier room.22 An entrance with anteroom is pre- served on the western side, while the northeastern corner is completely lost. Stairs connect with room 3 in which parts of the floor strata are preserved: an earlier brick floor beneath a later one in opus tessel- latum. Three bases for oil-presses were found iso-

The two structural parts of the villa were defined by Schwalb as "winter and summer residences," the peristyle nucleus for winter and the northern elon- gated wing for summer.24 Entrance is through the central part of the north wing of the peristyle, and it consists of a protyron projecting from the wall surface on four columns. Behind it, the cella ostiaria is divided only by a thin partition wall. The corri-

O 5 10 15

14

-9 ,, 13 2 I I8 10

12 i i 11

Ill. 4. Verige Bay-North, plan of villa

lated west of the cistern. A row of column bases was visible in the water in Gnirs' time, in front of the longitudinal wall. A narrow passage near 12 led from the thermal portico to the villa. Room 15 was probably added later, and, to the east, corridor 9 leads to the thermal section 8. In the northern part of that room there is a square basin with a semicir- cular back, while a black-and-white mosaic covers the floor. The water flowed from the cistern through room 7, where it was probably heated. In room 8 there is a smaller basin, 1 m. higher in level, built in opus signinum. The rest is lost in the sea. 3. BARBARIGA (porticusvilla): This villa was ex- plored by Hans Schwalb in 1901.23 Its nucleus is roughly 35 m. wide, and 26 m. long, and part of it was eroded by the sea (ill. 5).

dor opens to the courtyard around which Schwalb reconstructed columns, although he found no re- mains of them.25 All rooms in the north wing have simple floor mosaics in white, with black edging, with the exception of room C (ill. 4), where a much more elaborate polychrome effect is achieved by a black-and-white edging around a field in which a brown two-strand guilloche forms octagons in yel- low, red, brown and green. In the middle of this composition, a rich floral motif constitutes the visual and compositional center. Remains of marble revet- ment were found on the lower parts of the walls of this room, as well as fragments of architectural decoration.26

Room G, with a central position in the peristyle section, has a floor mosaic in four colors. Other

22 Gnirs (supra n. 13) 147. 23 H. Schwalb, Rdmische Villa bei Pola (Schriften der Balkan-

komission, Antiquarische Abteilung 2, Vienna 1902); some of the plates are reproduced in Mlakar, Die Rdmer (supra n. 7) pl. 6.

24 Schwalb (supra n. 23) 11. 25 Schwalb (supra n. 23) 15.

26 For opinions on the function of this room, see A. Gnirs, "Fruihe christliche Kultanlagen im siidlichen Istrien," Kunsthis- torische Jahrbuch 5, Beiblatt 4 (1911) fig. 2, and B.

Marugi(', "KrgSanstvo i poganstvo Istre u IV i V stolje'u (Le christianisme et le paganisme sur le sol de l'Istrie aux IVe et Ve sidcles)," ArchVest 29 (1978) 556, fig. 5.

58 ROBERT MATIJASIC [AJA 86

rooms have simple white floor mosaics with black

edging.27 A great number of sherds, mainly of am- phorae, were found in the corner room between the

4 A G IE i

0 5 10 15

Ill. 5. Barbariga, plan of villa

east and south wings. The row of rooms of the south wing opening onto the portico around the courtyard probably is residential; Schwalb recon- structed an upper floor in this part of the villa.28 Two prominent rooms (P) could have been storage magazines.

A narrow passage (H), connecting the peristyle with the thermal section, was originally a room of the eastern wing. The door to K was made later, only to be closed again towards the end of the life in the villa, as was the door between K and P. In the

thermal section, Schwalb identified two rooms with niches as apodyterium and caldarium (M). A pas- sage from corridor K leads to the cryptoporticus L and the promenade consisting of two wings forming an L-shaped structure (ill. 5, inset).

The northern section ("winter residence") is poor- ly preserved, in that only longitudinal walls have survived, but the cistern (today under water) is well preserved.29

The "porticusvilla" in Barbariga is similar in its conception to the Verige Bay villa. Schwalb's iden- tification of the "winter and summer residences"30

(ill. 5) is based on the location of the coast line which determined the placement of all structures to the west. Many elements-the floor mosaics, the attractive setting, a promenade section, the arrange- ment of rooms around a central courtyard/peri- style-suggest that the villa was purely residential. It may have been the center of an extensive estate, perhaps including the big olive-oil plant nearby. 3a. BARBARIGA-OIL PLANT: is north of the so- called peristyle villa (porticusvilla), and was exca- vated in 1953-1954 by S. Mlakar of the Archaeo- logical Museum in Pula.31 It was never published, but was mentioned by B. Maruli*32 (ill. 6). The central part is an elongated area containing oil-press

, ... .. O O0 0 c J6

bg4

Ill. 6. Barbariga, plan of oil plant

platforms, seven of which are preserved in a space which allows for the placement of as many as 12.

Each platform held two presses, so the manu- facturing facility had a large capacity, suitable for a plant in the richest olive growing area of the coast. Along the row of bases a drain carved in stone blocks led the oil to room b, adjoining to the west, in which there were two rows of square stone con- tainers. In a later phase, these were replaced by

27 Schwalb (supra n. 23) 17, p. 6. 28 Schwalb (supra n. 23) 19, p. 14. See also Mlakar, Die

Romer (supra n. 7) pl. 6. 29 Schwalb (supra n. 23) pl. 10 has reproductions of roof tiles

with stamps (QCLODIAMBROSI, AFAESONIAF, C. IVL. AFRICANI) as well as stamped amphorae (LIG, CAES, PON). The wall paintings he attributes to the third Pompeian style.

30 Schwalb (supra n. 23) 11. 31 Mlakar, Die Romer (supra n. 7) pl. 7. 32 B. Maruli', "Neki problemi kasnoantifke Istre u svjetlu ar-

heolo'kih izvora (Einige Probleme des spitantiken und byzantin- ischen Istriens im Lichte archiologischen Quellen)," Jadranski zbornik 9 (1973-75) 340.

1982] ROMAN RURAL ARCHITECTURE 59

circular ones. A millstone for olives was found in the southern adjoining room c, which was thus the first area involved in the process. The olives were

brought into one of the neighboring rooms to be milled into a mashy substance free of kernels. Un- der the presses, oil was extracted, left in basins for sedimentation, and then stored in dolia or packed in amphorae. The eastern part of the excavated area has four square basins for sedimentation (d), a facil- ity partly disturbed by a modern road.33

The Barbariga structure represents the third type of Roman villa (an agricultural factory), in being a plant completely equipped for the processing of ol- ives, ideally located in the fertile environs (ill. 6). From the ground plan alone, and in the absence of the final publication of the site, differential use of the space cannot be established, but the plant should probably be regarded as an important factory em- ploying workers and/or slaves. The existence of the plant is an affirmation of the quality and quantity of the Histrian oil distributed to Italy and the neighboring provinces.

Maruli6 has determined the date and duration of use of this structure. The earliest phase can be placed in the 1st century A.C., but the plant contin- ued in use for a much longer time.34 Many visible alterations in the building suggest that the manufac- turing section was reduced in size after several cen- turies, but it remained functional perhaps even after the 6th century A.C.35 4. VALBANDON: This villa is situated on Valban- don Bay, near Favana. It belongs to the category of spacious villas with luxurious decoration, although it is not of major importance. It is not certain whether the two groups of rooms to the north and south of the bay represent two villas or two parts of the same villa, which had to be divided because of the bay (ill. 7).36

The northern group of rooms is defined by two apses (ill. 8) (diameter 12 m.), both with floor mo- saics. In the neighboring room D, the earlier opus signinum floor was replaced with marble slabs. From here a small antechamber with a pentagonal floor mosaic (black-and-white pattern) leads to the procoeton (B). The lower part of the wall is covered with green marble, and the floor mosaic divides the room visually into two sections. Hall C (diameter

8.8 m.) is entered by means of several steps. Rooms C and H were later altered in size and shape, and probably became part of a peristyle court.

A

o. 200m

Ill. 7. Valbandon, plan of villa

0 5 10

H C

61

Ill. 8. Valbandon, N rooms, plan

The southern group of rooms (ill. 9) has three peristyles (A,K,F) with columns 2.38 m. high, crowned with Tuscan capitals. The northern peri- style has a central pool, built in a combination of opus signinum and opus spicatum. Rooms D,E,G and H have floor mosaics, and together with L and B, they surround peristyle A. In rooms D and E, 1st century A.C. coins were found (Augustus and Clau- dius). The mosaics were restored twice in later cen-

33 Mlakar, Die R6mer (supra n. 7) pl. 7. 34 Maru'i' (supra n. 32) 340. 35 Maruic' (supra n. 32) 340.

36 A. Gnirs, "Forschungen in Istrien," OJh 14 (1911) 155-69, see also Gnirs, "Grabungen in sudlichen Istrien," OJh 15 (1912- 13) 5-16.

60 ROBERT MATIJASIC [AJA 86

turies. Late 4th century coins were found around rooms B,C and F (Constantius I, Valentinianus

I).37

5

4

\\

S5 10 15

\ \

\C

A\E

L A

Ill. 9. Valbandon, S rooms, plan

The northern part of the southern group of rooms consists of two buildings. Building 1 is the continu- ation of rooms surrounding the peristyles A and K (ill. 9),38 but any connection which might have ex- isted was destroyed before Gnirs' excavations. Far- ther east are two water storage tanks on different levels, built in strong opus signinum and opus spica- tum. A strong wall (ill. 9,5) attached to the north- east of the water tanks may have been the support for a water conduit fed by a nearby spring. Coins of Augustus, Maxentius, Constantius and Theodosius were were found in this area, as well as a fragment

of sculpture-the left shoulder of a life sized toga- tus.

The villa in Valbandon probably had a character similar to that of the Barbariga villa, but it is unfor- tunately only partially preserved (ills. 7,8,9). The water-tanks, which may have been part of a con- siderable manufacturing establishment, suggest, however, some differences. In the south wing of the villa we recognize two square courts, one of which at least was surrounded by columns and had a pool in the center.39 The smaller court is surrounded by rooms paved with mosaics. In the Valbandon villa, the classical symmetrical scheme is obvious: two ar- eas with similar plans and functions (the atrium opening from the court, the peristyle from the gar- den)40 define an axial scheme with expansions on either side. Around the progression of atrium-tabli- num-peristyle are, among others, thermal areas with apses, magazines, cubicula and rooms for ser- vants and slaves.41 5. SIJANA: The villa is on the forested edge of a plateau northeast of Pula.42 The courtyard (C) is the central part, with rooms around three sides (ill. 10). The fourth side is closed by a wall. The resi- dential section, with the triclinium (D), extends into the northwestern part of the villa with the wheat storage room (F) and the stable (E) on one side of it. Protruding from the surface of the wall is the substructure of a tower (G), divided into two smal- ler rooms. On the other side of the residential quar- ters is a cowshed with storeroom (H), and room B (perhaps a slaves' bedroom).

Gnirs thought that the main entrance to the villa was through the opening in the southeastern wall, although vestibule A provides for communication between the stables, the residential section and the outside. He also remarked that room F had a floor in opus spicatum, which was raised one Roman foot above the level of the other rooms in the villa. Many fragments of amphorae and roof-tiles were found.43 6. RADEKI: The site is about 2 km. from the village of Loborika (Lavarigo), in the area of the Roman Municipium Nesactium. Its ground plan was drawn in 1906 (ill. 11), and published in 1908.44 As usual, the cistern is the best preserved

17 Gnirs, "Forschungen" (supra n. 36) 169. He makes no com- ments on the stratigraphic locations of the coins, so now we can only give a vague approximation of the period(s) of habitation.

38 Gnirs, "Grabungen" (supra n. 36) 5. 39 Sui6 (supra n. 1) 220. 40 Mansuelli (supra n. 1) 19.

41 See M. Vasii, "Rimische Villas vom Typus der Villa Ru- stica auf Jugoslawische Boden," Archaeologia lugoslavica 11 (1970) 62.

42 Gnirs (supra n. 13) 131-34. 43 Gnirs (supra n. 13) 133. 44 Gnirs (supra n. 13) 127.

1982] ROMAN RURAL ARCHITECTURE 61

element, and, judging by the remains of the sur- rounding walls, it was placed in the middle of the court. Built in opus signinum, its floor is covered

and glass. Fragments of stone capitals and columns, as well as a sundial, were also found. In the cistern, coins of Claudius and Constantine the Great were recovered.45 7. BANJOLE: The villa is visible today under the vegetation at the end of Paltana Bay, southwest of the village of Banjole (Balneolum), south of Pula (ill. 12).46 The front facing the sea is washed away,

Go 5 10

F E J

D'

A H B Ill. 10. Sijana, plan of villa

Ii

ii

If i I

1 r 0 5 10

Ill. 11. Radeki, plan of villa

with opus spicatum. The capacity was 80 m.3 of rainwater, which drained from the roofs into the impluvium. The residential area is on the northern side, and here Gnirs found many black and white mosaic tesserae. The extent and size of the walls can be followed around the cistern, yielding a ground plan similar to those of ?ijana and Kolci Hill, with the difference that the residential section is on the northern side. Among the finds, Gnirs mentions fragments of amphorae, dolia, oil-lamps

D E 5

A B C

J K L

._"• ....... • , o

Ill. 12. Banjole, plan of villa

but the rear part of the construction is preserved. Two units can be distinguished in the ground plan: D,E,F,G and H which is connected with J,K,L and M by passage A,B and C. The cistern was sepa- rated from the body of the villa (ill. 12, Z), and is today below sea level. It was built in opus signinum and is similar in its position to a cistern found near Barbariga.47

Rooms D,E,F,G, and H were probably the man- ufacturing section. An oil-press was found in room G, with a drain leading to the adjoining room (H) where a basin for sedimentation was obvious despite its destroyed state. It is probable that room F con- tained a wine press (oil presses usually occur in groups of three or more) and a basin for must. Room D could have been a storage room, and court- yard C a court or stable. Beneath room C a drain led from the factory to the sea. Only foundation walls are preserved of rooms J,K,L and M. 8. VELIKA ?ARAJA: The villa was excavated in 1956, by S. Mlakar, and a ground plan was pro- vided by M. Suit,48 but a comprehensive report has not yet been published. From a manuscript report49 we learn that the entire perimeter of the villa was explored, bringing to light a square ground plan

45 Gnirs (supra n. 13) 128, also Gnirs (supra n. 19) 148. 46 Gnirs (supra n. 13) 157-64. 47 Schwalb (supra n. 23) 20.

48 Sui' (supra n. 1) 221. 49 Manuscript report to the Archaeological Museum of Istra in

Pula, no. 195/1956, dated May 25, 1956.

62 ROBERT MATIJASIC [AJA 86

roughly 46 m. long, containing a partially damaged cistern and oil-press (ill. 13). The villa is situated near the village of Peroj, north of Fazana, on a slightly downsloping terrain (east to west). On three sides of the central court (D) are groups of rooms with manufacturing and residential functions. Pro- truding from the square is room H, which Mlakar identifies as a kiln or storeroom because it was out- side the main complex, although attached to it (see infra). A well built cistern (G), with three pillars, was placed on the edge of the courtyard. Near it is the storeroom for tools and implements, while sec- tion E is the residential part of the villa; it includes an exedra.

The northern section was designed for olive oil production. It includes anterooms (B) for the col- lection of olives, a set of rooms with the mill and presses, and four basins for sedimentation. A nar- row corridor linked the residential and manufactur- ing sections of the villa. Part of the foundation walls were destroyed.

Residential villas including factories (mostly for olive oil production) are therefore well represented in Histria. Of the five villas (Sijana, Kolci, Radeki, Banjole, Saraja), only three preserve the complete ground plan (Kolci, Sijana, Saraja). One is only partially explored (Radeki), while one is almost completely destroyed (Banjole). The focal point of these villas was a courtyard, a peristyle, or a portico on the front, with combinations of architectural ele- ments depending on the functions.50 Of the Histrian villas, those in Sijana (ill. 10) and Saraja (ill. 13) had central courts, as perhaps did that in Radeki (ill. 11), while the villa on Kolci Hill had a peri- style (ill. 1). The residential and manufacturing sec- tions of these villas are almost indistinguishable ex- cept for some details of the ground plan. It is proba- ble that such villas were run by individual farmers, rather than by Roman citizens living in town, as were the more luxurious villas: Verige Bay (ills. 2,4,7), Barbariga (ill. 5) and Valbandon (ills. 5,6,7), which were used mainly as holiday retreats. In in- dustrial villas, the position of the residence was sub- ordinate to the factory, as in the villas on Kolci Hill, in Sijana and in Saraja.

Four presses were placed in a room of the villa

on Kolci Hill, which Vitruvius would consider rath- er inadequate: Ipsum autem torcular ... ne minus

longum pedem XL constituatur. Latitudo eius ne minus pedum senum denum ... 51 The villa in

Saraja has the presses on the north side, against Vitruvius' advice. In both cases there are two or

SG e

10 0F 0 1

I11. 13. ?araja, plan of villa

three rooms added to the villa for the storage of olives or for heating the water necessary in the ex- traction of oil.52 Room H (ill. 13), projecting from the southwestern corner of the villa in Saraja, may have been a storeroom for fodder or a drying kiln: Horrea, fenilia, fararia, pistrina, extra villam facien- da videntur, ut ab ignis periculo sint villae tuti- ores.53 Room C in ?araja may have had the same function, in which case the adjoining room could have been a stable.54

With the strengthening of the frontiers of the Empire on the Danube in the 1st century A.C., the Histrian peninsula received greater protection from the raids of nomadic tribes; thus began the rise of its economic prosperity. In addition to the development of the Colonia Pola, the ager was divided into even squares (centuriatio) and allotted to colonists. A

50 Suit (supra n. 1) 220. 51 Vitr., De archit. 6.3. 52 Vitr., De archit. 6.2, 6.3.

53 Vitr., De archit. 6.5. 54 Vitr., De archit. 6.2, 7.4.

1982] ROMAN RURAL ARCHITECTURE 63

great number of Roman rural architectural remains show the vigor of the area from the 1st to the 3rd centuries A.C. (ill. 15). Lack of elaboration in the

publications before 1914 makes the precise dating of artifacts more difficult.55 In most cases, however, numismatic finds can be used to estimate the span of time during which a particular villa was in use.56

If we divide the territory of Colonia lulia Pola into three equal zones corresponding to the distance from the western coast,"57 we notice that the density of Roman rural remains varies. There is a concen- tration of villas on the sea coast and in the imme- diate hinterland, their number declining the farther one moves from the sea (ill. 14). Unfortunately, the villas on the coast were submerged by the rising sea level during the past 2000 years, and the sea has

destroyed large portions of them. Some parts of the ager had been left unallotted

(ager insolutus), remaining as public pasture. Such was the case with the southern tip of the peninsula (Premantura-Promontorium), the northeastern part of the ager facing the frontier of Provincia Dalma- tia, and the Brioni Islands.58 The fact that the land was not allotted does not mean that there were no Roman farms in these areas; but they were not

officially colonized, although architectural and other remains exist sporadically.

The density of rural sites is greatest between Val- bandon and Barbariga, on the coast and immediate- ly inland. Between Barbariga and Rovinj-Rugi- nium, there are two concentrations: around Colone- Sv. Pavle (St. Paul) Bay, and in the Vistar (Vi- strum) region, both on the coast. The Medulin and Premantura peninsulae south of Pula were also set- tled. The suburbs of Pola were also populated, as witnessed by the remains of suburban villas of lux- urious character.59

Of the coastal villas, all except Barbariga are on a bay (Verige, Valbandon, Banjole), because access from the sea is easier, and the view more pictur- esque. The response of the architecture to the set- ting is most evident in villas with sea views, where

the fronts are articulated by porticoes and peristyles. More modest villas with simple fronts have enclosed central courts or atria. The architects and planners

PARENTIUM -,

a

..... /

.8

POLA

,

. I

"* ,i

I .1

.

Ill. 14. Distribution of villas in Colonia lulia Pola

of Roman rural villas seem to have preferred down- ward slopes or terrace-like gradients, with the villa reaching down to sea level, a feature characteristic of other villas along the eastern shore of the Adri- atic Sea.6o Villas in the hinterland are often sited in similar positions, on slopes or terraces (Sijana, Ra- deki, Saraja, Kolci Hill). The owner or builder con- sciously chose less fertile pieces of land on which to

55Almost all the excavated artifacts were lost or destroyed during the two World Wars, when the material was moved frequently. Some of them may still be deposited in Italian Museums.

56 See Vasi6 (supra n. 41) 45. For the evidence, Gnirs (supra n. 13) 128; Gnirs, "Forschungen" (supra n. 36) 169; Gnirs, "Grab- ungen" (supra n. 36) 9.

57 See Sui6 (supra n. 1) 212. 58 Sui6 (supra n. 1) 96-104; Sui6, "Ostaci limitacije nagih pri-

morskih gradova u ranom srednjem vijeku (Remains of the Limi-

tation of our Towns in the Early Middle Ages)," Starohrvatska prosvjeta n.s. 5 (1956); Sui6, Limitacija gradova na istocnoj jad- ranskoj obali (The Limitation of Towns on the eastern Shore of the Adriatic Sea) (Zbornik Instituta za historijske znanosti, Za- dar 1955); R. Chevallier, "La centuriazione romana dell'Istria e della Dalmazia," Attilstr n.s. 9 (1961) 11-24.

59 Mlakar, Die Ramer (supra n. 7) 14. 60 Sui6 (supra n. 1) 217, M. Zaninovi6, "Neki primjeri smje?-

taja anticwkih gospodarskih zgrada na obalno-ototkom podrubju Dalmacije," Arheololki radovi i rasprave JAZU 4-5 (1967) 359.

64 ROBERT MATIJAMIC [AJA 86

build the center of the farm, because of his depen- dence on the availability of abundant, fertile soil.61

SITE: c 0 (0 JC

C" c (S( E) ,4 C 0C C

M . ELEMENT: > > - > )

a

mosaic 1 U 1 1

fresco 1

capital 1 O stucco 1

peristyle I O r-

yard I I U I

thermae I I cistern I I U I

oil-press O U U I U U

wine-press O r

storage-roorr M M M

stabling I U U

ELEMENT: CO =O

c.C c CC.

ATATI I I I I

/DATATION:I " ,. c_

u, u

Ill. 15. Villas in Colonia lulia Pola, contents and dates

Although the rural villa included everything nec- essary for the independent existence of a familia rustica, it was nonetheless linked in some ways to the nearest town. Economic relationships with the closest colonies produced an improvement in Hi-

strian rural life from the 3rd century onwards. The break up of the large-scale trading networks, roads and fortifications, the incursions of northern tribes across the Danube and Rhine, caused the larger factories to become reduced gradually to small pro- duction units not dependent on the markets of Italy, Noricum and Pannonia. Although manufacturing activities still continued on most villa sites, the villa as an institution-as a part of the total Roman economy-did not survive. The local production and distribution system which thus developed was not affected by the crises of the Late Empire, and many villas in Histria survived until a much later date, even as late as the first Slavic raids at the end of the 6th century A.C.62

This was the case with the Barbariga oil-plant, which may have been subject to radical restrictions in the capacity of production (additional partition walls reduced the number of presses to three or four) but some presses were used until the 6th cen- tury A.C.63 The concentration of villas at some points along the coast was so great that they must have formed the rural nuclei of many Medieval villages. A large number of place-names ending in -anus (changed to -an, -ana) scattered along the western coast of Histria bear further witness to the continuity of life and settlement in the dark cen- turies between the fall of the Roman Empire and the formation of independent towns in the 11th- 12th centuries.64 The role of the Christian faith in this process in Histria is not completely known, but some of the chapels and churches in the ager are known to have been built on Roman villas, as they were in other parts of the western provinces of the Empire.65

ARHEOLOSKI MUZEJ ISTRE M. BALOTE 3

52200 PULA, YUGOSLAVIA

61 Zaninovii (supra n. 61) 359, 369; Percival (supra n. 1) 157-61.

62 Percival (supra n. 1) 169-71; Maruli6, "Slavensko-avarski napadi na Istru u svijetlu arheologke gradje," Peristil 2 (1957) 63-69.

63 Marusic (supra n. 32) 340. 64 Suic (supra n. 1) 223-24; C. De Franceschi, "La topono-

mastica dell'agro polese," Attilstr 51-52 (1939-1940) 119-97; Percival (supra n. 1) 31-33, 172-78.

65 Percival (supra n. 1) 32-33, 183-99, B. Schiavuzzi, "Attra- verso l'agro colonico di Pola," Altilstr 24 (1908) 91-171; Ma- ruvi', "Spomenici sakralne arkitekture s upisanom apsidom," Histria Archaeologica 5 (1974) 14, 26, 32, 49-53, 61-65. I am grateful to the personnel of the Archaeological Museum in Pula, particularly to Prof. Stefan Mlakar. My thanks go also to Prof. Rendie Miocevic, under whose direction this article was conceived.