• 1 For the final report of the American excavations at Seleucia, see Hopkins 1972.

2During the last century, the site of Seleucia —about 30 km south of modern Baghdad— was the object of intense investigations which saw two different archaeological missions following one after another: the American expedition of the University of Michigan and the Italian expedition of the Centro Ricerche Archeologiche e Scavi di Torino per il Medio Oriente e l’Asia (fig. 1). The excavations were started at the end of the 1920s by American archaeologists, who conducted six campaigns between 1927 and 1937. Their efforts were mainly focused on dwelling block G6, located south of the wide ship canal which cuts across the city more or less lengthwise. The excavations, which reached the Seleucid levels only in extremely limited parts, led to the complete exploration of the Parthian phases of the block.1

• 2 All the information given in the present paper on the elements of architectural decoration found by (...)
• 3 The American excavations recognized four levels of occupation (Hopkins 1972, p. 6): Level IV (307-1 (...)

3From dwelling block G6 come the large majority of the roughly 200 elements of architectural decoration brought to light by the American excavators.2 As shown in figure 2a, there were basically two materials employed in the architectural decoration of the housing insula: terracotta and stucco. As illustrated in figure 2b, their distribution varies considerably through time. Terracotta elements are most extensive in Level III of the American excavations, which corresponds to the first Parthian phase of the city, and significantly decrease in number in Levels II-I, corresponding to full and late Parthian times.3

• 4 Goldman 1972, p. 128.

4According to the American archaeologists, there is no clear evidence attesting the use of stucco prior to the Level II phase, as the fragments found in Level III might be out of context.4 As a consequence, the introduction of stucco in the architectural decoration of the block appears to be contemporaneous to the evolution of domestic architecture, from a house with a two-columned portico on one side of the courtyard to a house with an iwan, a rectangular space which is entirely open on the short side overlooking the courtyard.

• 5 Goldman 1972, p. 132-134.

5From Level III of the block come a dozen fragments of terracotta simas, along with some roofing tiles and several antefixes of clearly Greek inspiration. A number of them were found in Street 6 and in rooms opening on it and, according to Goldman, probably belonged to the roofline decoration of the street façade, while others were brought to light in courtyards and in their surrounding rooms. No fragments of simas were found in Level I, and even if three of them come from Level II, their original provenance is dubious.5 Apart from the roofline, the houses of Level III were most restrained in their architectural decoration: only a few rooms contained some elements such as decorated bricks or miniature capitals.

• 6 Goldman 1972, p. 143-144.
• 7 Goldman 1972, p. 137-142. On the stuccoes from the American excavations at Seleucia, see also Debev (...)

6However, some of the rooms of Level II show traces of a rich stucco ornamentation, which decorated arches, door jambs, ceilings and walls. The main entrance to the block was richly decorated ; moreover, stucco elements were concentrated in the courtyards and in the rooms opening to them.6 If terracotta elements are of genuinely Western inspiration and use many of its standard motifs —such as volutes, palmettes, vine grapes, bead and reel, egg and dart— the stucco decoration is characterized by the use of all-over patterns based on the repetition of simple geometrical forms —circles, parts of circles and squares— interlaced to produce rosettes, petals and diamond motifs.7

• 8 For the report of the excavations on the western side of the agora, see Messina 2006; for the easte (...)
• 9 For the final report of the excavations at Tell ʿUmar, see Messina 2010.
• 10 For the excavation report of the South Square area, see Graziosi 1968-1969, Negro Ponzi 1970-1971, (...)

7In some respects the materials from the Italian excavations broaden the picture offered by the architectural fragments from the American expedition. The Italian archaeologists conducted fourteen campaigns on the field: eleven between 1964 and 1976 and three between 1985 and 1989. The main areas investigated were the northern agora (the so-called Archives square), where, in Parthian times, dwelling-houses and handicraft activities were built on the remains of Seleucid public buildings;8 Tell ʿUmar, the great artificial mound located on the far northern end of the urban perimeter, which conceals the remains of the civic theatre;9 and the so-called “South Square”, a vast open space not foreseen in the original urban layout and situated on the southern limit of the city, close to a large commercial street.10 From these areas come almost all of the elements of architectural decoration brought to light.

• 11 For the simas from the American excavations, showing the same decorative pattern, see Goldman 1972, (...)
• 12 See Mallwitz, Schmidt & Wetzel 1957, p. 47-48, Taf. 32ac.
• 13 See Vlassopoulou 1990, p. xxiv-xxv, n. 44-45, from Athens (eave tiles decorated with a painted “oli (...)
• 14 For an overview on the Greek roof decoration in Hellenistic times, both in marble and terracotta, s (...)
• 15 See Kose 1998, p. 233, Taf. 114, 1-5.

9Conversely, many of the terracotta elements are similar to the ones from the American excavations. This is the case of the roofline decorations, which were particularly abundant at Tell ʿUmar and in the South Square area. They include 29 antefixes, at least 40 fragments of sima and a couple of tiles and cover tiles. In the great majority of the sima fragments, the face is framed by plain mouldings and bear a triple-leaf decoration running between continuous dots or a bead and reel border; the lion-headed spouts are usually very worn (fig. 3). In the more detailed specimens, unfortunately very fragmentary, thin stems bearing buds (or berries) enrich the leaf motif, while in the more stylized versions the leaf design is simplified (fig. 4): stems disappear and the clusters are formed by three leaves either arranged in a fan-like shape or, less frequently, running parallel or almost so to each other.11 Apart from the actual simas, characterized by an L-shaped profile because of the presence of an extended heel on the back, at Seleucia the same motif occurs on flat terracotta plaques of more or less the same size. In Mesopotamia, the leaf and berry motif occurs on some stucco fragments from Babylon.12 In the Greek world, the olive branch commonly appears on the painted decoration of classical architectural terracottas (simas and eave tiles),13 whilst in Hellenistic times terracotta simas are more commonly decorated by spirals and vegetal scrolls.14 Closer to this latter decoration are three fragments from the South Square, bearing volutes and stems ending with rosettes (fig. 5). The face of these latter fragments is slightly higher (around 15 cm), whereas the height of the simas with triple-leaf design ranges between approximately 9 and 11 cm. A similar pattern —spirals on the sides of a lion-headed spout— is attested on some fragments of simas from Uruk, found in the Irigal area.15

• 16 See Valtz 1988, p. 25-26, fig. 16.
• 17 We are talking about examples Inv. S3815, found in square CLXXI, 64ae (h. 20.8, w. 16.4 cm), and In (...)
• 18 See Mallwitz, Schmidt & Wetzel 1957, Taf. 23, c-d, from Babil.

10Antefixes are relatively tall: the height of the complete specimens ranges from 21 to 26 cm approximately. The palmette is composed of leaves with upturned tips emerging from a spade-shaped centre, and the base is usually decorated by horizontally spreading volutes or stems. One of the best preserved specimens comes from the Seleucid levels of the eastern side of the Archives square, and is dated to the end of the 3^rd cent. bc (fig. 6).16 From the South Square area come two complete examples most likely from the same mould, one of which was found together with its relevant covering tile (fig. 7).17 Similar antefixes —with volutes at the base of the palmettes, composed of leaves with upturned tips— were found in the area of the Summer Palace at Babylon.18

• 19 See Negro Ponzi 1970-1971, p. 35.

11As mentioned above, the great majority of the elements of roofline decoration come from Tell ʿUmar and the South Square. About half of the fragments from Tell ʿUmar come from the area of the temple leaning on the western front of the theatre: however, we cannot positively state that they originally belonged to the decoration of the temple itself, as they were mainly found in the fill layers that sealed the structure after its abandonment or in the Seleucid layers preceding its first building phase. The vast majority of the fragments of simas from the South Square area were dispersed over several rooms of the large building located on the western side of the square, and a few antefixes were found in the street flanking the western front of the building itself: according to the preliminary report, these probably decorated the front of the structure in levels IV or even Va, corresponding to Seleucid and early Parthian times.19

• 20 See Negro Ponzi 1970-1971, p. 35-36.
• 21 See Pierpont Morgan Library, Morgan Seal 594 (a bull attached by two winged lion griffins; the scen (...)
• 22 For an overview on the use of the griffin motif in Hellenistic architectural sculpture, see Webb 19 (...)
• 23 See Goldman 1972, p. 143.
• 24 See Keall, Leveque & Willson 1980, p. 26, fig. 15, 1.
• 25 See Goldman 1972, p. 143, fig. 73.
• 26 See Kose 1998, p. 322, 360, Taf. 114,7, 115,1-4, 123,5.

12From the late Parthian levels of the same building (level I of the excavations) comes a backed brick figured frieze found in the large room 64bf (fig. 8); from the same room also come the above-mentioned stucco capital (fig. 9) and small stucco fragments, suggesting that the walls of the room were covered with a rich decoration. Other fragments of terracotta and stucco come from the adjacent rooms, giving the impression that the entire aisle was richly ornamented.20 The terracotta frieze, in two fragments, is hand-cut and topped by stucco mouldings of western inspiration. The figured scene, framed by a palmette, represents a humpbacked bull attacked by two winged griffins. The subject is of ancient Near-Eastern tradition : griffins appear on Mesopotamian cylinder seals from the 2^nd millennium bc, and the winged lion-griffin attacking a bull is well attested on seals of Middle Assyrian period.21 In Hellenistic times, the griffin appears as an architectural motif both in the Near East and the Mediterranean area.22 The report on the architectural elements from the American excavations of Seleucia quotes several fragments of bricks showing a griffin attacking a deer,23 whilst the motif of the humpbacked bull between griffins is attested on a stucco frieze from Qalʿeh-i Yazdigird.24 A brick from the western side of the Archives square represents a griffin flanking a vegetal element (fig. 10), a motif also attested among the findings of the American expedition.25 At Uruk, bricks with lions and griffins decorated the pilaster-capitals of the Gareus temple, and are also attested among the decorations of the so-called Partische villa.26

• 27 See Goldman 1972, p. 131-132, fig. 38-40.
• 28 See Kose 1998, p. 361, Taf. 126,1-2 (from Uruk); Legrain 1930, p. 39, n. 442, pl. LXXVI (from Nippu (...)

13The Italian excavations also yielded some fragments of terracotta cornices and decorated bricks (around 40 in total), again mainly coming from Tell ʿUmar and the South Square area. Some of the cornices bear multiple plain mouldings, whilst other are enriched with standard Greek patterns such as the bead and reel or the egg and dart. The bricks, which were used for continuous friezes on cornices and capitals, were decorated with vegetal motifs such as scrolls, volutes and palmettes. Some of them are decorated with a simplified version of the grape and leaf design, with stems and tendrils reduced to a curved line that encircles the single elements, thereby isolating one from another (fig. 11). The motif, which also recurs on some terracotta and stucco fragments from the American excavations,27 is very popular in Parthian Mesopotamia. At Uruk, it is attested on some stucco and terracotta fragments from the so-called Partische Villa, and is also present at Nippur, Babylon, Assur and Hatra.28

• 29 For the complete catalogue of these elements, see Menegazzi 2014, p. 791-801, tav. 718-732.
• 30 For the excavation report of the sounding, see Menegazzi 2009. For the activity of the workshop, se (...)

14Antefixes, simas, decorated bricks, fragments of cornices and friezes were made of low fired clay mixed with straw, basically the same mixture used for making bricks, and represent approximately 40 % of the terracotta elements of architectural decoration brought to light by the Italian archaeologists. The remaining 60 % consists of elements for assembled capitals, capitals and bases for engaged pillars, the great majority of which come from a small sounding located on the southern side of the Archives square.29 They were concentrated in two deposits, linked with the activity of a large workshop specialized in the production of terracotta figurines: in the same area —of about 214 m²— over 3 000 figurine fragments were found, either lying in pits or in accumulation layers.30 As a matter of fact, elements for assembled capitals, capitals and bases for engaged pillars were made in the same fine and compact clay used for making the terracotta figurines.

• 31 See Goldman 1972, p. 129.
• 32 See Menegazzi 2014, p. 801, tav. 732.
• 33 For a careful description of the fragments of assembled capitals from Tell ʿUmar, see Invernizzi 19 (...)
• 34 See Invernizzi & Lippolis 2008, p. 248-257.

16As previously mentioned, elements for assembled capitals come almost exclusively from the deposits on the southern side of the Archives square. Only one fragment of acanthus leaf is recorded among the American finds from block G6,31 and just five specimens (one volute and four leaves) were found in the South Square area.32 Based on these observations, we can hypothesize that this kind of decoration was not meant for domestic architecture. It might not be by chance that the only fragments of assembled capitals come from the area of Tell ʿUmar. The two best preserved ones show that acanthus leaves and volutes were fixed with clay to a tile that was applied to the masonry core of the capital by means of gypsum mortar. The finishing of the assembled capitals consisted of a thin veil of blue-green glaze, which concealed the joins and strengthened the cohesion of the individual elements.33 Assembled capitals apparently have no parallels in Mesopotamia ; conversely, as is well known, they find a quite close parallel in Old Nisa.34

• 35 See Kose 1998, p. 362, Taf. 127-129 (from Uruk); Kröger 1982, Taf. 23,4 (from Larsa); Andrae & Lenz (...)
• 36 See Menegazzi 2014, p. 708-709, n. 20.G148-G149, 20.G163, 20.G165, tav. 641.
• 37 See Invernizzi 1995, p. 5, fig. 6.

18Like the assembled capitals, these latter were most likely meant as a revetment for mud brick masonry. Corinthian figured capitals or half-capitals made of stucco were quite popular in Parthian Mesopotamia, being attested at Assur, Larsa, Uruk, and among the materials from the American excavations at Seleucia.35 Scale models of columns and pillars with figured capitals are also attested among the terracotta figurines from Seleucia.36 By contrast, capitals made up of tiny terracotta plaques seem to be a production unique to Seleucia: once again, the closest parallel can be traced outside Mesopotamia, at Nisa, where Ionic capitals made of single relief plaques are attested.37

• 38 See Goldman 1972, p. 134.
• 39 See Valtz 1988, p. 25-26.
• 40 See Martinez-Sève 2002, p. 39-44. Elements of roofline decoration, together with tiles and cover ti (...)
• 41 According to Strabo (XVI, 1-5) at Seleucia “they do not use tiles for their roofs, because there ar (...)
• 42 A. Invernizzi, personal communication.

19To sum up, elements of architectural decoration shed some light on the complex processes of interaction and exchange between Greek and local cultures in a field, that of architecture, deeply influenced by geographical factors, such as the well-known lack of building stones in Mesopotamia. Elements of roofline decoration seem to imply the presence, at Seleucia, of Greek-style sloped roofs both in housing and in public areas,38 as suggested by the finds from dwelling block G6, from the South Square and from Tell ʿUmar. On the eastern side of the Archives square, fragments of architectural decoration —including the antefix reproduced in fig. 5— were concentrated in a carefully planned dwelling house located east of the Stoa: they were found laying on the floor of two rooms, together with a large amount of burned remains of wooden beams.39 An analogous find context is recorded at Susa, in the area of the Ville Rojale: antefixes and fragments of simas were found together with fragments of burned beams among the remains of a peristyle house built around the middle-second half of the 3^rd cent. bc and reported to have been covered with a double sloped roof.40 On the other hand, it has to be pointed out that the elements of roofline decoration from Seleucia are very few if compared to the whole extent of the excavated areas. As a consequence sloped roofs, if present, must have covered only some of the buildings of Seleucid and early Parthian times, and maybe only limited parts of them. Moreover, the extreme scarcity of tiles and cover tiles brought to light by the excavations at Seleucia has to be taken into consideration.41 In the light of this latter observation, we cannot rule out the possibility, pointed out by Antonio Invernizzi42, that Greek elements could also be used for the decoration of flat roofs.