1The excavations revealed a part of the ancient town including a NE/SW orientated street and at least two insulae, one on each side (Fig. 4). There are two main phases of occupation, both ending with a destruction of the buildings. The layout of the habitation dates to the late Classical/early Hellenistic period, in which the town, both residential zones and sanctuaries, was reorganized. After the destruction of the buildings from the second phase of occupation in the late Republican period, the area was frequented in a sporadic manner during Imperial times.

II.1 The phases and general stratigraphical analysis

2Phase 1: Early Hellenistic period

3The first one covers a period from the second half of the 4^th to the early 3^rd century BC and is best preserved in the eastern insula. Here the Courtyard House (trenches H, J and L) and the building in trench G (perhaps being part of this house) were provided with ashlar blocks for the socle courses, some exceeding 1.0 m in length, while mudbrick formed the upper walls supported by timber posts. The Courtyard House was decorated with polychrome stucco, including floral motifs, and probably also bicolour stucco (red and white) and of architectural type. The courtyard was paved with large stones. Some of the roof tiles were painted. In J V two pantiles with completely preserved width were found. The ridge of the roof was probably aligned NE/SW, otherwise more fragments should have been located in trenches H and L.

4Phase 2: Late Republican period

5The second phase seems to have begun sometime in the 2^nd century BC. It ended during the first half of the 1^st century BC at the latest. This phase is only represented in the western insula. In general the walls were built of smaller blocks and reused material. There is documentation for bicolour stucco (red and white) and murals of architectural type. Floors consisted of terra battuta (stamped earth with e.g. chips of pottery and small stones), clay and reused roof tiles. The latter is located in a room in trench E which served as a kitchen and / or service room. This interpretation is suggested by the presence of a drain in the room and the huge amounts of pottery found. The ridge of the roof in this insula has not been located/detected.

6Phase 3: End of Republican period-early Imperial period

7Various activities seem to have taken place in the later 1^st century BC/early 1^st century AD including a renewal of the street.

8Phase 4: Imperial period

9After these periods the area was still frequented and sporadic occupation might have taken place. It also seems to have functioned as a dump site. There are a few later walls, but their time of construction cannot be precisely dated. In the 3^rd century AD the northern part of the area was used for child burials. As far as the abandonment of the site is concerned, the material is clear: except for a few items, the latest finds (African Red Slip Ware, glass, lamps, coins) are datable in the 4^th to 5^th centuries AD.

10General stratigraphical analysis

11For an overview of the status of the trenches and strata, Tables la-1b and 2 should be consulted. Here the broad chronological framework of the individual trenches and strata is presented together with the indication of bulldozed strata, location of the bedrock etc. Some trenches were subdivided and excavated in separate sections, meaning that the general consecutive numbering of the layers does not always follow the vertical stratigraphy; for instance strata III and IV are used in the same horizontal layer where they designate the western and eastern part of the trench respectively. For the trenches in question, additional tables must be consulted. This applies especially to two of the trenches initiated in 1986 (D and G) when only trial trenches were dug, and the stratigraphy of the area was totally unknown. During the excavation of these trial trenches, however, it soon became obvious that all material in strata I (modern humus layer) and II was mixed and contained modern material. Subsequently, these strata were removed by bulldozer in the 1989- and 1990-seasons, and they are not described in the overview of the trenches and their stratigraphy (chapter II. 1.1). The next layer, stratum III, is relatively homogenous in the whole excavated area. It is characterized by a loose and sandy soil of brown to greyish brown colour. In the northern part (trenches H, J and L) this stratum was disturbed in ancient times by child burials. As to the other strata and their correlation, Tables la-1b, 2 and the stratigraphical analysis should be consulted, since the same stratum designation may represent chronologically different phases.

12The travertine bedrock was reached at various points in the excavated area, both to the east and west of the street; in trench A, Ds, F, G and N. In the latter (stratum IX) the surface was quite uneven with levels varying from -709 to -772 in the four corners and the deepest measurement being -786. This also applies to trench F where the points vary from -710/719 (stratum VII) to -777 (strata X and XI). In the remaining trenches the bedrock was reached at the following levels: A VII-735/738; Ds XI -760/767; G XV -797/798. In some cases the bedrock was worked to receive the lowest courses of the wall (socle courses).

13Material before the layout

14The earliest finds from the Prop. Avallone excavations are early Iron Age and Archaic sherds of the 7^th to 6^th centuries BC (impasto, bucchero and argilla depurata), not associated with any structures. The extreme fragmentation and concentration in trench N strata III-II testify to their secondary context. The episodic presence of these sherds in other trenches and strata is by no means surprising. Material of the late Archaic and Classical periods comprises blackgloss kylikes (e.g. K VI), an Attic red-figure sherd (Ds XI) and two coins belonging to the earliest bronze issues of Poseidonia (cat. nos. LL-6 and LL-7). The coins are dated to the end of the5^th century BC and around 400 BC.

15Phase 1: Early Hellenistic period: First street

16The first building activity is represented by the terra battuta street (-785/788) and fill beneath in G XIV+XVI-XVII. As this street was cut/dug through by the foundation trench in G XV, it gives a terminus post quem for the layout of the insulae, but whether there were years, months or days between the two events cannot be answered. Datable finds consist of a few black-gloss fragments, one with Gnathian-type decoration. They only give a broad chronological framework of the later part of the 4^th and early 3^rd century BC.

17Phase 1: Early Hellenistic period: Layout of insulae and the second street, occupation and destruction

18The time of construction is dated to the second half of the 4^th century BC by the finds from the foundation trench in G XV (e.g. cat. no. 0-36, Morel series 2433). Judging from the massive collapse material in trench G and the Courtyard House (trenches HL), the destruction took place at the latest in the early 3^rd century BC. The street belonging to this phase is documented in G XII (-738/740) and Ds XI (-748), the lowest part. Finds from the destruction layers include, for example South Italian red-figure, Morel series 2423, 4311 and 4382, reticulate lekythoi/flasks, core-formed glass (cat. no. DD-1) and three anepigraphic obols (cat. nos. LL-1 – LL-3).

19The remains from this phase are located in: (Fig. 61)

20Trench A, stratum VII: Wall Al? (tl -713/735.5). Cleaned up after the destruction.

21Trench Ds, stratum XI: partly the lowest course of wall D1 (bl -748). Cleaned up after the destruction.

22Trench E, strata X-IX: wall El (tl -760/72, bl -793/799). Cleaned up after the destruction.

23Trench F, stratum XI: probably part of wall FI (bl for southern part -748/754). Cleaned up after the destruction.

24Trench G, strata XV, XIII-VIII (below terra battuta): stratum XV = foundation trench, stratum XIII = fill below terra battuta street in stratum XII at -738/740, strata XI-VIII (below terra battuta) = collapse material fallen upon the street. Wall G1 (tl -670 in stratum VII, bl -797/798).

25Trenches H-L, strata VI-IV (lower part): Courtyard House, courtyard with top level -661/663 in H V, paved area? north of wall J2 in J V (-684/706).

26Phase 2: Late Republican period: Rebuilding west of the street, occupation and destruction

27During the period of abandonment a layer of lime sediment was created in trench E (bottom of stratum VIII). This does not necessarily mean, however, that the zone was abandoned for a long period, as the geological conditions in the area favour such sediments1. In trench Ds the street received a new layer of terra battuta. Its northern continuation is located in E VIII.

28Based on the coin (cat. no. LL-10, dated 211-210 BC) found in Ds XI below the terra battuta floor (-750), the rebuilding took place after the end of the 3^rd century BC. Apparently, the insula east of the street was abandoned. The collapse material, primarily from Dn VI + Ds X-IX and E V, dates the time of destruction within the first half of the 1^st century BC at the latest. The destruction layers contained much black-gloss ware (e.g. Morel series 2122, 2822, 2973-2974), thin-walled ware (for instance the situlashaped beaker with and without barbotine dot decoration), common ware casseroles types A and D, but hardly any glass and no terra sigillata sherds.

29The remains from this phase are located in: (Fig. 62)

30Trench A, stratum VI: There are no structures, but the collapse material found here is parallel to the one from trenches D and E. The layer is disturbed and mixed with finds of Imperial date.

31Trench B, strata VII-IV (lower part): Walls B1 and B2. Terra battuta or clay floors in strata VI+VII (-733/738 and-750/752). Stratum VIII is not datable.

32Trench C, strata VI-V: Walls C1 and C2.

33Trench Dn, stratum VI + Ds, stratum XI/X-IX: Wall D1 was reused, new wall D2 and terra battuta/clay floor (-731/749/751); renewal of terra battuta street.

34Trench E, strata VIII-V-IV (lower part): Wall E2 was constructed on top of wall El, L-shaped wall and tile pavement (-710.6/726). Remains of terra battuta located between level -741/752 and-751/756 are probably part of the NE/SW orientated street.

35Trench F, stratum XI/X-V-IV (lower part): Wall FI was reused, new wall F2.

36Material visible in the southern balk in Ds VIII was interpreted as lapilli (between -653 and-668). If this is correct, it gives a terminus ante quem for the destruction. Apparently the area was afterwards abandoned/unoccupied.

37Phase 3: End of Republican period-early Imperial period: Third street

38The collapse material in G VIII was sealed by the new terra battura street at -678/695. Stretches of this street are also documented in the easternmost parts of trenches E and F. The demolition of wall E2 and the building of wall E3 (lining of stones) probably took place in relation to this renewal/alteration. Probably these alterations were made in the later part of the 1^st century BC/early Imperial period. This dating, tentative only, is indicated by the absence of African Red Slip Ware, which is present in the successive strata (E IV and F IV). Three African Red Slip Ware sherds derive from the upper part of G VIII (Hayes Form 27 and two unassigned). These, however, should be regarded as intrusive, as a late Republican coin (cat. no. LL-8) was found in G VII, revealing the disturbed/complex situation here. Wall A2 in A V was perhaps constructed at this time judging from the mixed character of stratum VI below (both Hellenistic and Imperial material).

39The remains are located in: (Fig. 63)

40Trench A, stratum V(-III?): Wall A2.

41Trench Dn, stratum V + Ds, stratum VIII.

42Trench E, stratum V (upper part): Terra battutta street (-679/680) and stone lining (wall E3, -670/679), wall E2 was pulled down.

43Trench F, stratum V (upper part): Terra battuta street (-679).

44Trench G, strata VIII(upper part)-VII: Terra battuta street (-678/695) in VIII.

45Phase 4: Imperial period: Frequentation, walls and burials

46The whole zone was frequented in the Imperial period, 1^st to 4^th/5^th centuries AD and used for various purposes.

47New walls were constructed in trench M IV, and perhaps in A V-III, thus indicating occupation on the outskirts, whereas the amount of collapsed material from Ds V and G VI-V (no structures) in the central zone points to the usage as a dump site. The northern part of the area was used for burials of children (trenches H, J and L) in the 3^rd century AD. Moreover, there were skeletal remains of a teenager in H IV, and a fourth child burial is possible in F IV. The strata are characterized by terra sigillata, African Red Slip Ware (forms of the 1^st to the 3^rd centuries being predominant), common ware casserole type E, glass (the majority dating from the 2^nd century AD onwards) and lamps with short, round nozzle and a perline/a globetti. The latest ancient material dates from the5^th to the 7^th centuries AD.

48The remains are located in:

49Trench A, strata (V?)-III: Wall A2.

50Trench B, strata IV(uppermost part)-III.

51Trench C, strata IV-III.

52Trench Dn, strata IV-II + Ds, strata VII-III.

53Trench E, strata IV-III (cf. Table 2 for the complex character of the layers).

54Trench F, strata IV-III (cf. Table 2 for the complex character of the layers).

55Trench G, strata VI-III (cf. Tables 2-3 for the complex character of the layers).

56Trench H, strata IV(upper part)-III.

57Trench J, strata IV-III.

58Trench K, strata IV(upper part)-III.

59Trench L, strata III.

60Trench M, strata IV-III.

II.1.1 Trenches A-N and stratigraphy

61The trenches and structures in Prop. Avallone are aligned NE/SW. For convenience of description, e.g. when relating the structures to each other, a conventional north is used here, corresponding to northeast. Levels are often given with the highest and lowest separated by a slash (/). Only the most significant finds or groups of material are listed in order to give an impression of the character and date of the individual trenches and strata. As far as pottery is concerned, references are made to typologies, both established by others (Morel 1981, Conspectus 1990, Hayes 1972) and by the contributors in the present publication (common wares, coarse wares, lamps).

Trench A

62Excavated 1990, strata I-VII (Figs. 5-6).

63Trench A (Quadrante 187 and 173) is L-shaped and covers the area within the coordinates 77.5-81.0/116.5-119.5 (not comprising the area within 77.5-79.0/116.5-118.0). To the east of wall A2 (c.80.10-81.0/116.5-119.5) the excavation stopped with stratum III in the southern half (116.5-118.0) and with stratum IV in the northern half (118.0-119.5). West of wall A2 the excavation stopped in stratum III in the southern half (79.0-c.79.5/116.5-118.0) and continued to stratum VII in the area 77.5-C.79.5/118.0-119.5. Here the travertine bedrock was reached at level -735/738.

64Phase 1: Early Hellenistic period (?)

65Stratum VII: The double wall, wall Al, aligned NW/SE, together with the burnt clay floor and two postholes almost on alignment represent the earliest use of the area (Fig. 6). No foundation trench was recovered; so the time of construction cannot be ascertained. The burnt clay floor testifies to the destruction of the building, and since no collapse material was found directly on the structure, the area must have been cleaned up. There are, however, some finds suggesting that the structure was in use in the early Hellenistic period.

66FINDS

67A VII: black-gloss ware (Morel series 2787), black-gloss ware with overpainted decoration; unguentarium (cat. no. S-2) with high marked shoulder and relatively tall stem; common ware casserole (type A = Hellenistic/Republican).

68Phases 2-3: Late Republican period-imperial period

69Stratum VI is defined as a destruction layer, which does not seem to be associated with any of the excavated walls in the trench. The layer was rich in finds of iron, e.g. nails and various lumps. Lead objects were also found.

70The chronological range of the finds is wide and could suggest that the layer has been jumbled up. Both late Hellenistic/Republican (relief ware) and Imperial (Italian Sigillata) material came from the upper part of the destruction layer. Below the destruction layer a lamp (cat. no. EE-14) of the 1^st century AD was found. The disturbance might be related to the construction of wall A2 (cf. below).

71Stratum V: Wall A2 (with tl at -571 in stratum III) is documented as far down as -674 (stucco in situ on its western side), i.e. immediately above the collapse material in stratum VI. Its relation to wall Al is not clear, but perhaps it was constructed some time after the destruction ending the second phase of occupation, and this would have resulted in disturbance of the collapse material that had not been cleared away. The time of construction cannot be dated precisely. The layer contained e.g. black-gloss ware and glass, and taking into consideration the finds of early Imperial date in stratum VI, the wall might go back to the late 1^st BC or 1^st century AD. It is not clear whether the wall continued in use throughout later centuries when new courses might have been built (cf. below), or whether it remained visible in a ruinous state during these centuries. The latter eventuality would represent frequentation of the area and not occupation.

72FINDS

73A VI: roof tiles; black-gloss ware; red-gloss ware; relief ware (Long Petal Bowl type, c. 150-75 BC); thin-walled ware base with relief stripes; Italian Sigillata (Consp.26 variant); common ware lids with top and orlo annerito (type B2); glass unguentarium; open lamp of Greek type (type Al), lamp a volute (type B2b = 1^st cent. AD); iron nails.

74A V: black-gloss ware; glass.

75Phase 4: Imperial period

76Strata IV-III: Wall A2 is attested in strata IV-III (tl -571) and has a gap at the centre; to the east a concentration of blocks seems to constitute a (curved?) NW/SE wall (wall A3) (tl -588). Whether this wall continues further down is not known. Both strata IV and III contained terra sigillata forms of the 1^st century AD, and in stratum IV a coin of the late 3^rd century was found at level -615, immediately to the west of wall A2. In stratum III large pieces of tiles were located at the northern balk (79.0/119.5) and in the corners at 79.0/116.5 (-587/597) and 77.5/118.0 (-584).

77FINDS

78A IV: floor tiles; Italian Sigillata (Consp.28); common ware lids orlo annerito (type B3); amphora fragments (unassigned); glass; bronze coin (cat. no. LL-12: Claudius II Gothicus, 270 AD or shortly after); bronze fibula.

79A III: floor tiles; Pompeian-Red Ware (late 1^st cent. BCearly 2^nd/3^rd AD); Italian Sigillata (Consp.23); African Red Slip Ware (unassigned); common ware casserole (type E = late 2^nd-4^th cent. AD); amphora fragments (unassigned); glass (unassigned); lamp a perline/globetti (type B5d = second half of 1^st to 3^rd cent. AD); lead weight.

Trench B

80Excavated 1990, strata I-VIII (Figs. 7-8).

81The trench is rectangular (Quadrante 188) covering the area within the coordinates 76.75-79.5/130.0-131.5. The deepest strata, VII and VIII, were excavated in isolated areas in the trench.

82Phase 2: Late Republican period

83Strata VIII-IV (lower part) represent the occupation and destruction of the building (walls B1 and B2). Wall B1 is documented in strata V-IV (lower part) and wall B2 in stratum V. The collapse material, concentrated in the eastern part of the trench, defines stratum V-IV (lower part) and was found lying on clayish layers, interpreted as floors, in strata VI (east of wall B2, -733/738), and VII (the area between walls B1 and B2, -750/752) (Fig. 8). Large quantities of pottery were found. Stratum VIII represents an isolated area of a darker colour within stratum VII (77.0-c. 78.3/131.2-131.5) (Fig. 8). Reused material was employed in wall Bl.

84The material points clearly to destruction of the site before the Imperial period, in the first half of the 1^st century BC.

85FINDS

86B VII: common ware beaker (Hellenistic/Republican).

87B VI: black-gloss ware; iron stud.

88B V: black-and brown-gloss ware (Morel 1312/1523, 1315, 1534, 2122, 2790, 2788, 2822), black-gloss ware with overpainted decoration (Morel 2978); red-gloss ware; unguentaria; thin-walled ware (one fragment with barbotine dot decoration); common ware casserole (type A = Hellenistic/Republican, type B = c.275-150 BC), platters orlo bifido (type 2/5b = late 2^nd cent. BC-lst cent. AD), lids with top (types Al & A2); coarse ware clibanus and mortarium; Graeco-Italic amphora (Hellenistic) and amphora fragments (unassigned); biconical lamp (type A2b = Hellenistic).

89B IV (lower part): the finds from B IV are listed together, see below.

90Phases 3-4: End of Republican period-imperial period

91Strata IV (uppermost part)-III. Stratum IV represents a mixed layer with material from the destruction and random material from the frequentation of the area afterwards. Stratum III belongs to the same phase of frequentation.

92FINDS

93B IV: black-gloss ware (Morel 1313, 2111, 2671, 2822, 2961, 2973-74/2977-2978); terra sigillata; glass; common ware platters; iron nail.

94B III: Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD, strata II-III).

Trench C

95Excavated 1990, strata I-VI (Figs. 7 and 9).

96The trench is rectangular (Quadrante 188) and covers the area within the coordinates 76.75-79.5/132.5-135.5.

97Phase 2: Late Republican period

98Strata VI-V represent the occupation and destruction of the building (walls Cl tl -670/673 and C2 tl 677/682, large worked limestone block tl -688/693). Stratum VI was excavated in an isolated area (76.75-77.5/134.0-135.5), i.e. the north-western part of the trench, where a canal was revealed (-697/704), and in the north-eastern part (77.75-79.5/134.0-135.5). A large amount of stucco was found in the north-eastern corner of the trench. Stratum V is defined as a destruction layer comparable to the one in B V-IV (lower part). The function of the large block and the canal is not clear; nor the relation between these features and the walls. The material belongs to the late Republican period, 2^nd to mid-1^st century BC. The uppermost part of stratum V was mixed with finds of Imperial date.

99FINDS

100C VI: white stucco; black-gloss ware (Morel 2433, 2981), black-gloss ware with overpainted decoration; redgloss ware; common ware casserole (type A = Hellenistic/Republican); coarse ware dolium (type F2); loomweights.

101C V: black-gloss ware (Morel 1552, 1700, 2700, 2944), black-gloss ware with overpainted decoration (figurative); red-gloss ware (Morel 2671, 2788); thin-walled ware (e.g. situla beaker with barbotine dot decoration); common ware casserole (type A = Hellenistic/Republican); coarse ware (e.g. louterion); iron stud.

102C V (uppermost part): Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD); African Red Slip Ware (unassigned); common ware casserole (type E = late 2^nd-4^th cent. AD).

103Phases 3-4: End of Republican period-imperial period

104Strata IV-III represent a period of frequentation of the area from the end of the Republican period to the Imperial period after the destruction of the building. Material from the uppermost part of stratum V belongs to this phase as well. The content of lime flecks and pottery fragments increased at the bottom of stratum IV.

105FINDS

106C IV: black-gloss ware; Italian Sigillata (stamped fragment); African Red Slip Ware (Hayes Form 27); common ware casserole (type A = Hellenistic/Republican); lamp with short round nozzle (type B4 = c.50-early 2^nd cent. AD).

107C III: African Red Slip Ware (unassigned, strata II-III).

Trench D

108Excavated 1986, 1989-90, strata I-XI (Figs. 10-21).

109The trench is rectangular (Quadrante 174) comprising the area within 80.5-84.0/122.0-127.5 (without 80.5-81.0/122.0-125.0). In the northern part, designated trench Dn (80.5-84.0/125.0-127.5), the excavations reached stratum VI, while in the south, trench Ds, it proceeded until stratum XI where the travertine bedrock was reached both to the east of the wall (-761/765) and to the west (-760/767).

110Concordance of the strata:

111Phase 1: Early Hellenistic period

112Stratum XI is defined as the layer below the terra battuta/clay floor west of wall Dl, i.e. inside the building. East of wall Dl the stratum was excavated in the northern part only in the area not covered by the terra battuta street (Figs. 14-15 and 19). The bedrock was reached, thus the stratum must go back to the initial phase, but at the same time it also seems to be the foundation layer for the second phase of occupation after the area had been cleaned up. The earliest part of wall Dl is formed by the lowest course (bl at -748) consisting of large ashlar blocks in limestone located in Ds. The bottom level for wall Dl is at -748 in 83.17/123.7. No finds are directly associated either with the time of construction or with the time of destruction. However, residual material from stratum XI inside the structure indicates an early Hellenistic date. The bronze coin, cat. no. LL-10, found at level -750 is dated to the end of the 3^rd century BC and might be residual material or it might have been dropped here in relation to the rebuilding. It is difficult to determine whether the triangular outlet in wall Dl belongs to this phase as well or the succeeding (Figs. 17-18). The lowest part of the terra battuta street (tl -744/751) should also go back to the original layout. In this part the layer contained larger pieces of tile than in the layer immediately above. The terra battuta covers the southern part and a narrow strip along the eastern balk in the northern part. Here in the area dug through collapse material was revealed at level -759, including e.g. large fragments of roof tiles. An irregular pit dug into the bedrock was also found here (bl -782/785). It was covered with reused roof tiles with traces of fire. The pit is aligned with the outlet. The pit probably represents an alteration or repair during the second phase of occupation or in relation to the rebuilding in phase 2. This would explain the cutting of the terra battuta street.

113FINDS

114Ds XI: black-gloss ware (Morel 2671); coarse ware dolium (type K); amphora fragments (unassigned), table amphora; loomweights; bronze coin (cat. no. LL-10: end of 3^rd cent. BC); iron studs.

115Phase 2: Late Republican period

116Strata XI/X-IX (trench Ds) + stratum VI (trench Dn) represent rebuilding, occupation and destruction. Judging from the coin found below the terra battuta inside the building, this rebuilding took place at the earliest at the end of the 3^rd century BC. Earlier collapse material was cleared away and wall D1 was reused: a reused block (a basin) was incorporated into the lowest course and constitutes the angle with the new transverse wall, wall D2. A new course, composed of small stones, was put on top of the earlier wall. The large flat stones forming the uppermost course of wall Dl, although with a slightly different orientation, also belong to this phase, as the surface at -676/677 (Ds IX) is in keeping with the surface level from the northern stretch of the wall, -676.5/686 (Dn VI). Moreover, collapse material was found above the surface of the wall in both areas. A bronze coin (cat. no. LL-15) was found in situ on the wall at level -686. A terra battuta/clay floor was made inside the building. Apparently, the terra battuta constituting the street to the east of the wall (level -744-747/745-751) was renewed, this time with smaller fragments of tile.

117Strata X-IX + VI are characterized by a massive destruction layer including e.g. stones and numerous, often large, fragments of roof tiles (Fig. 16). In Ds the destruction layer found to the east of wall D1 is located at a slightly lower level than the one west of the wall. The collapse material west of wall Dl contained e.g. stone blocks and stucco. Beneath a step (bl -733) located in the south-eastern comer at the junction of walls D1 and D2 a loomweight and vessel were found, the latter lying on the terra battuta at -739 (Figs. 20-21). The destruction layer, beginning at -666/669, from Dn was for the greater part left in situ (Fig. 16). Much of the pottery is secondarily burnt.

118No terra sigillata or glass finds were made in Ds IX and Dn VI, and the material is of late Republican date. Thus the destruction took place before the mid 1^st century BC.

119Unfortunately, two of the coins found in stratum X are unidentifiable. The third one, cat. no. LL-5, came from an area where the clay floor was probably dug through; it is thus more likely to represent residual material than forms part of the collapse material proper.

120FINDS

121Dn VI: roof tiles; stucco; black-gloss ware (Morel 1100); red-gloss ware; unguentaria; thin-walled ware (fragment with barbotine decoration); common ware casserole (type A = Hellenistic/Republican), lids orlo annerito (types B2 and B3); coarse ware dolia (e.g. type FI); lamp a volute (type B2a = Augustan to first third of 1^st cent. AD); iron stud.

122Ds X: roof tiles; stucco; black-gloss ware (Morel 1313, 1514, 1520, 1536, 1730, 2153, 2285, 2640/2650, 4372?), black-gloss ware with overpainted decoration; red-gloss ware; thin-walled ware (e.g. situla beaker found at level -739), common ware casserole (type A = Hellenistic/Republican), various jars, lid with top (type A2); amphora (Dressel 1A = c.130-50 BC, Dressel 20? = Tiberian/Claudian- 3^rd/4^th cent. AD); rounded lamp (type A2a); bronze coins (cat. no. LL-16: unidentifiable, cat. no. LL-5: 3^rd cent. BC); many iron finds (e.g. handles, sickle, studs, wall-fixture, slags).

123Ds IX: stones/blocks; roof tiles; stucco; mortar; blackgloss ware (Morel 1536?, 2671, 2770), black-gloss ware with overpainted decoration, brown-gloss ware; red-gloss ware; thin-walled ware (including fragments with barbotine decoration and cord undulations); common ware casserole (type A = Hellenistic/Republican, type D = late Republican), lids with top (type A2) and lids orlo annerito (type B2); cylindrical lamp (type A2c = Hellenistic); bronze coin (cat. no. LL-15: unidentifiable).

124Phase 3: End of Republican period-early Imperial period

125Dn V + Ds VIII are the layers immediately above the massive destruction layer, which is reflected by the increased number of finds in the lower part of the strata. The quantity of black-gloss sherds decreased considerably, and terra sigillata and glass occur for the first time. Thus the strata represent frequentation of the area in the late Republican period into the first 1^st century AD. Some residual material from the collapse material is also likely to be present, especially at the bottom of the layers.

126FINDS

127Dn V + Ds VIII: unguentarium; thin-walled ware (with rouletted decoration); Italian and South Gaulish Sigillata (Consp.28, 3.2/4.6/6/20-21); African Red Slip Ware (unassigned fragment); glass (tubular base); rounded lamp (type A2a = Hellenistic).

128Phase 4: Imperial period

129Dn TV + Ds VII represent frequentation in the area. In the northern part the layer was almost void of finds.

130Dn III+Ds V+VI. In this period of use/frequentation the number of African Red Slip Ware and glass increased considerably. A bronze coin of the mid-4^th century AD (cat. no. LL-13) was found at -603 in Dn III.

131Dn II (lower part)+Ds III+IV. In the southern part the layer was mixed with modern glass and nails.

132FINDS

133Dn IV+Ds VII: unguentarium; thin-walled ware (barbotine pine scale decoration); terra sigillata (unassigned); African Red Slip Ware (Hayes Forms 8B, 23B); lead-glazed ware (ribbed handle); glass (bowl with two handles, tubular rim); lamps (type B4 = c.50-150 AD and second half of 1^st-3^rd cent. AD).

134Dn III + Ds V+VI: Pompeian-Red Ware (end of 1^st cent. BC-beg. of 2^nd/3^rd cent. AD); Italian Sigillata (Consp.48.1?); African Red Slip Ware (Hayes Forms 3, 23B, 27, 61 transitional); common ware casserole (type E = late 2^nd-4^th cent. AD); glass (beaker with trailed decoration, bowl/beaker with wheel-cut geometric decoration dated 3^rd to 4^th cent. AD); lamp (type B4 = mid- 1^st to early 3^rd cent. AD, type B5b, type B5e with globuled decoration, type B7b = c.425-650 AD); coin (cat. no. LL-13: Constantius Gallus Caesar, 352-354/351-354 AD).

Trench E

135Excavated 1989-90, strata I-X (Figs. 22-33).

136This rectangular trench (Quadrante 174) comprises the area within 80.5-87.5/129.5-131.5.

137The excavations began in the western half of the trench (80.5-83.5/129.5-131.5). Later they were extended to the central (83.5-84.5/129.5-131.5) and the eastern (84.5-87.5/129.5-131.5) parts. Strata VIX was excavated in an isolated area in the central part of the trench between the tile pavement to the west and wall E3 to the east (c.82.5-84.5/129.5-131.5). In the eastern part the excavation stopped at -679/685.5 and-689/694 and in the west at the level of the tile pavement.

138Phase 1: Early Hellenistic period

139Strata X-IX represent the earliest use of the area with wall El, small stones and a pit in stratum IX. In the southern part a layer of lime sediment was located (tl -752/755). The excavation continued in stratum X in an. isolated area (Figs. 26-27, 29 and 32). The bedrock was apparently not reached. Wall El consists of two large blocks and a third one situated below the dressed block from strata VIII-VI; thus the wall continues towards the south.

140As part of the rebuilding of the structure in the succeeding phase the area was cleaned up. Consequently, the number of finds was very low, and no datable fragments were recovered.

141FINDS

142E IX: roof tiles; red-gloss ware.

143Phase 2: Late Republican period: Kitchen/service room

144Strata VIII-V-IV (lower part): Rebuilding of the structure, occupation and destruction. A new wall was constructed on top of the remains of the first one; the large block in limestone (wall E2, strata VIII-VI), situated partly below the pavement, is all that is left of it. A triangular outlet built of stones was located in the western balk between 131.33-131.5 (at the bottom) (tl -739; bl -754). Probably, it formed a drainage system with the cover tile found on the pavement inside the building (Fig. 33). The outlet must have continued through the later demolished wall.

145Strata VIII-VI represent mixed or disturbed layers; there may be collapse material from this phase as well as material from the time of the demolition of wall E2. Residual material from the first phase of occupation cannot be excluded either. During the rebuilding or demolition (cf. below) the layer of lime sediment, created during the period of abandonment, was cut. Stratum VIII was almost void of finds at the bottom. A terra battuta layer was registered in the northern part of stratum VIII; it probably represents a street whose level (between -741/752 and-751/756) corresponds to the bottom level for the outlet.

146While the eastern facade of the building is missing due to alteration at a later date, the tile pavement, composed of reused roof tiles, and L-shaped wall (wall E4, northern stretch appearing in stratum IV) from stratum V are still in situ. The surface of the floor is highly uneven (-710.5/726, c.81.25-83.1/129.5-131.5) (Figs. 28 and 30-31). Clearly the pavement has sunk in the central part, but not in the south-eastern comer where the large dressed block is situated underneath (Fig. 27). Large fragments of roof tiles (-709/717) were found on the pavement and large stones (-703/707) on the tiles, revealing the sequence of collapse with the roof first followed by the walls (Figs. 30-31).

147Judging from the levels of the lowest course of wall E4 (-728) and the pavement (-710.5/726), they seem to belong to the same phase.

148The finds from strata VIII-VI are often tiny pottery sherds. The time of destruction is given by the material from the massive destruction layer in stratum V, e.g. found on the tile pavement, namely in the first half of the 1^st century BC. Many of the pottery fragments were adjoining. The lower part of stratum IV had a higher content of black-gloss sherds and belongs to the collapse material from stratum V. The quantity of pottery and repertoire of classes together with the tile pavement and drainage leave no doubt that the room served as a kitchen or combined kitchen-service room with storing of both cooking and table wares.

149In the uppermost part of stratum V a singled-course wall/lining of stones (wall E3, tl -670/679) is located in the central part of the trench and in the east a terra battuta street (c.86.5-87.5/129.5-131.5, -679/680) (Figs. 24-25). These two features together with the demolition of wall E2 are likely to belong to an alteration after the destruction, probably in the late 1^st century BC or early 1^st century AD. The southern part of the street was dug through, and here fire-blackened stones appeared. Immediately to the west of the terra battuta there is a very hard layer, which probably forms part of the renewal of the area as well.

150FINDS

151E VIII: few roof tiles; black-gloss ware; coarse ware basin (type B).

152E VII: black-gloss ware (Morel 2787); common ware lids with top (types A1 and A), jars (types 2 and 4).

153E VI: black-gloss ware (Morel 1520, 1552, 2700), brown-gloss ware; red-gloss ware; common ware casserole (type A = Hellenistic/Republican).

154E V: stones; stucco; roof tiles; South Italian red-figure; much black-gloss ware (Morel 1313, 1315, 1520, 2111, 2122, 2284, 2285, 2720/2730, 2787, 2822, 2942, 2974, 2973-2974/2977-2978), black-gloss ware with overpainted decoration; red-gloss ware; unguentarium; thin-walled ware (with and without barbotine dot decoration-both beaker and jar); common ware casserole (type A = Hellenistic/Republican), platters orlo bifido (typel/5a = late 2^nd cent. BC-lst cent. AD), jar (types 2-4), lekythos, frying pan; coarse ware basin (types A-C), dolia, clibani, mortarium; amphora (Dressel 1A and unassigned), biconical lamp (type A2b = mid- 3^rd/later first half of 1^st cent. BC); terracotta statuettes; loomweight; iron (e.g. fibula, fishing hook, knife, wall-fixture).

155E IV (lower part): the finds from stratum IV are listed together, see below.

156Phases 3-4: End of Republican period-imperial period

157The lower part of strata IV belongs to stratum V (e.g. fragment from stratum IV adjoining fragments from stratum V), while the upper part represents frequentation in the Imperial period. The mixed chronology is clearly visible.

158Stratum III represents frequentation in the later Imperial period.

159FINDS

160E IV: black-gloss ware (Morel 2233, 2830), brown-gloss ware; thin-walled ware (with and without barbotine dot decoration); Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD); Italian and South Gaulish Sigillata (Consp.36.2-4 and unassigned); African Red Slip Ware (Hayes Forms 23B, 27); common ware casserole (type A = Hellenistic/Republican, type D = late Republican, type E = late 2^nd-4^th cent. AD), platters orlo bifido (type l/5a = late 2^nd cent. BC-lst cent. AD), lekythos; amphora (Dressel 2-4 = later 1^st cent. BC-mid-2^nd cent. AD); glass; lamp a volute (type B2b = 1^st cent. AD); terracotta statuette; iron (e.g. pair of pincers?, stud, nails, wall-fixture).

161E III: thin-walled ware; Italian Sigillata (Consp.34.2); common ware casserole (type E = late 2^nd-4^th cent. AD), lids orlo annerito (types B2 and B4); African Red Slip Ware (Hayes Forms 23B, 27); much glass (mainly parallels in the 2^nd cent. AD to c.300 AD); lamp a perline/a globetti with round nozzle (type B4 = late 1^sl- 2^nd cent. AD).

Trench F

162Excavated 1989-90, strata I-XI (Figs. 34-39).

163The trench is irregular (Quadrante 174) covering the area within the coordinates 80.5-82.5/137.5-138.5 and 82.5-86.5/136.5-139.5. Excavations began in the eastern part of the trench (84.5-86.5/136.5-139.5) which was extended towards the west (80.5-83.5/ 136.5-138.5) and later the balk (or the central part) 83.5-84.5/136.5-138.5 was excavated in order to reveal the NE/SW wall (wall FI). Strata VIII-XI were excavated in an isolated area west of wall FI, 82.5-83.4/137.0-138.0. The central part of the trench had been disturbed by modem plants as far down as stratum V.

164Phase 1: Early Hellenistic period

165Strata XI: The burnt soil and bones together with stones having traces of fire should date from the destruction that ended the first phase of occupation. The layer also contained a posthole (or pit) (tl -756; bl -777) (Figs. 38-39). The bedrock was reached at -752/759 and -760/777.

166Probably a part of wall FI also belongs to this phase, at least the large blocks in the lowest course of its southern half (Fig. 36). The bottom level of the wall in this part is -754.

167After the collapse material was cleared away.

168FINDS

169F XI: black-gloss ware (Morel 4382); red-gloss ware; common ware jar (type 4), krater (type 4); coarse ware mortarium (type A3); amphora (unassigned fragment).

170Phase 2: Late Republican period

171Strata XI/X-IV (lower part): Wall FI was rebuilt (tl -671 in stratum IV; bl -748/754 at the transition between strata X and XI) or reused and a new transverse wall (wall F2, tl -688, appearing in stratum VI) was constructed to the west. The latter is built on clayish soil and is composed of blocks of varying sizes, small stones and fragments of tiles (Fig. 37). A posthole is situated in the northern part of wall F1 and immediately to the west there is another one.

172The collapse material from strata VI-IV (lower part) belongs to this second phase of occupation, which ended with the destruction in the first half of the 1^st century BC. A possible floor (hard and compact layer) is located in stratum VIII south of wall F at level -712/715. In the northern part the bedrock appeared in stratum VII at -710 and -719 (82.5/139.5 and 83.4/139.5).

173The finds from stratum XI/X-VIII should represent the phase of rebuilding, which disturbed the area. It might therefore contain residual material from the first phase of occupation.

174The collapse material was concentrated in strata VI-IV (lower part). In stratum VI it consisted almost exclusively of roof tiles. In stratum V the material was found primarily in the northern part and west of wall FI. One terra sigillata sherd came from stratum V. However, it is clearly intrusive, as it was found in the area disturbed by modern plants. Stratum IV contained mixed material, however, with a clear tendency for the Hellenistic/Republican finds to come from the lowest part of the stratum, thus belonging to the collapse material in stratum V, and for the Imperial finds to come from its upper part. For finds in this stratum, cf. below.

175The terra battuta-paved area in stratum V in the easternmost part of the trench (c.85.1-86.5/137.5-139.5) at -679 is identified as a street (Fig. 35). South of this area the terra battuta was dug through (84.5-86.5/136.5-137.5, below-682). The stones in this part were blackened, probably due to fire. Pieces of charcoal were also found. This street layer should be dated to a period after the destruction (identical with the situation in trench E).

176FINDS

177F X: roof tiles; black-gloss ware; common ware jar (type 4).

178F IX: roof tiles; black-gloss ware; common ware jar (type 4).

179F VIII: roof tiles; black-gloss ware; common ware casserole (type A = Hellenistic/Republican), jar (types 2 and 4); coarse ware (unassigned dolia, situla).

180F VII: black-gloss ware (Morel 2424); red-gloss ware; brown-gloss ware (Morel 2985c, 5226); common ware casserole (type A = Hellenistic/Republican), krater (type 3), jar (type 4).

181F VI: large amount of roof tiles; black-gloss ware (Morel 1552, 4260, 4311), brown-gloss ware; red-gloss ware (Morel 2985); Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD); common ware jar (types 2, 4 and 8); loomweight.

182F V: roof tiles; stucco; mortar; black-gloss ware (Morel 1315, 1323, 1520, 1552, 2424, 2985), brown-gloss ware; thin-walled ware (beaker); terra sigillata (unassigned); common ware casserole (type A = Hellenistic/Republican, type D = late Republican), carinated bowl, jar (type 2), lid with top (type A6); loomweight; iron (e.g. fibula, studs, wall-fixture).

183Phases 3-4: End of Republican period-imperial period

184Stratum IV: The lower part belongs to stratum V. Here the destruction layer (west: -657/662, including e.g. large fragments of tiles and large stones) and the surface of wall FI appeared (-671). The upper part represents frequentation of the area in the later Imperial period.

185Uppermost part of stratum IV: Remains that may be interpreted as a child burial ad enchytrimos destroyed in antiquity were found at the eastern balk (-637/646).

186Stratum III represents frequentation.

187FINDS

188F IV: roof tiles; black-gloss ware (e.g with stamped rosette); red-gloss ware; terra sigillata; African Red Slip Ware (Hayes Form 23B); common ware casserole (type A = Hellenstic/Republican), jar (type 2), lids with top (types A2, A6 and A); coarse ware basins (type B), dolia (unassigned), situla; glass (one fragment approximately dated Neronian/Flavian to 3^rd cent. AD); loomweight; metal (iron, bronze).

189F III: thin-walled ware (with rouletted decoration); terra sigillata; African Red Slip Ware (Hayes Forms 8B, 50?); common ware casserole (type E = late 2^nd-4^th cent. AD), jar (type 1), lid with top (type Al), lids orlo annerito (types B3 and B5); glass; metal (bronze ring).

Trench G

190Excavated 1986, 1989-90, strata I-XVII (Figs. 24-25 and 40-44).

191This rectangular trench (Quadrante 174) comprises the area within 85.0-91.5/126.0-128.0. The excavation began in the western part (85.0-88.0) and was later extended to the east (88.0-91.5). The trench was excavated in a very complicated way. For the relationship between the strata in the different parts of the trench, see Table 3.

192Phase 1: Early Hellenistic period

193Strata XIV + XVI and XVII represent the earliest activity in the area when the first street paved with terra battura was made (-785/788). Strata XIV + XVI are identical with the street (tl -785/788, bl -797.5/799) and stratum XVII the layer of soil beneath (layer 13 on Fig. 41). Here it is possible to see that wall G1 is located on top of a worked part of the bedrock with top level at -797/798. In stratum XVII the bedrock was reached at -800 in the southern part.

194The finds point to a date in the second half of the 4^th century BC.

195Strata XV (foundation trench, tl -789) and XIII (earth fill)-XII (terra battuta street) represent the time of construction of the building (wall G1, tl -670 in stratum VII) and the second terra battuta street (-738/740). The foundation trench, c.25.0 cm wide, cut the older street in stratum XIV + XVI (Figs. 40-42; layers 11-12 on Fig. 41). Here a drinking service composed of an oinochoe and a concave-convex cup were found. Both had been deliberately cut in half (Figs. 43-44)2.

196Strata XI-VIII (below the terra battuta) represent the period of occupation and the time of destruction. The terra battuta in stratum VIII was dug through in the west to level -699, left in situ in the area c.86.0-88.0, and again removed further to the east except for some very hard and compact zones (cf. Figs. 24-25). Below this terra battuta collapse material appeared (especially tiles, pottery and larger stones). The surface of the massive destruction layer was located at -679, and its material was removed, including e.g. two bronze coins (cat. nos. LL-7 and LL-9).

197In stratum XIII an outlet in the wall appeared. Much charcoal was observed along the northern balk, especially in the area 88.5-89.2 in stratum XI. In stratum IX a layer of ashes appeared at -724.

198Judging from the finds from the foundation trench, the time of construction can be dated in the second half of the 4^th century BC. The material from the destruction layers (strata XI-VIII below terra battuta) belongs to the late 4^th and 3^rd century BC. The coin, cat. no. LL-9, found in stratum VIII at level -692, is, however, dated to the end of the 3^rd century BC and might have ended up in the upper part of the collapse material in relation to frequentation of the area after the first destruction or when the layer was disturbed during the making of the third terra battuta street. The other coin, cat. no. LL-7, from stratum VIII is even earlier (around 400 BC?) and from a slightly higher level (-686).

199FINDS

200G XVII: black-gloss ware, black-gloss ware with Gnathian-type decoration.

201G XVI: black-gloss ware.

202G XIV: black-gloss ware (Morel 2570/2580); common and coarse wares.

203G XV: black-gloss ware: foundation deposit composed of a concave-convex cup (cat. no. 0-36, Morel 2433) and an oinochoe (cat. no. 0-88).

204G XIII: common ware casserole (type A = Hellenistic/Republican).

205G XII: black-gloss ware.

206G XI: roof tiles; black-gloss ware; common ware jar (types 2 and 4).

207G X: roof tiles, mudbricks; black-gloss ware (Morel 2283, 2787, 2789, 4311); red-gloss ware; common ware casserole (type A = Hellenistic/Republican), jar (types 2 and 5), krater (type 2), lid with top (type A7); coarse ware dolia (unassigned), mortarium (type H), situla; iron pin.

208G IX: large stones and fragments of roof tiles; mudbricks; black-gloss ware (Morel 2285, 2423, 2424, 2787), black-gloss ware with overpainted decoration, brown-gloss ware; reticulate lekythoi/flasks; red-gloss ware; common ware casserole (type A = Hellenistic/Republican), lids with top (type A); amphorae (unassigned); rounded lamp (type A2 = Hellenistic); terracotta statuette; metal.

209G VIII (below terra battuta): many roof tiles, mudbricks; black-gloss ware (Morel 1765?, 2251, 2284, 2423, 2973-2974/2977-2978), black-gloss with overpainted decoration; common ware casserole (type A = Hellenistic/Republican), lid with top (type A); coins (cat. no. LL-7: Poseidonia around 400 BC(?), cat. no. LL-9: end 3^rd cent. BC); bronze spatula and two unidentified fragments.

210Phase 3: End of Republican period-early Imperial period

211Stratum VIII (above terra battuta)-VII: The third street, in stratum VIII at -678/695, sealed the destruction layer of the first phase of occupation. The material from the upper part of stratum VIII belongs to stratum VII, and the coin from this stratum (cat. no. LL-8, level c.-670) dated to the late Republican period shows that the area to the east of the wall was frequented in this period. In stratum VII wall G1 was uncovered from 126.0-128.0.

212FINDS

213G VIII (above terra battuta): African Red Slip Ware (Hayes Form 27).

214G VII: red-gloss ware (pinkish-purple pre-sigillata); terra sigillata; common ware lid with top (type A2); coin (cat. no. LL-8: late Republican).

215Phase 4: Imperial period

216Strata VI-V + IV + IVa represent a layer of collapse material from the Imperial period, the majority belonging to the 2^nd century AD. No structures are associated with this phase, but the many roof tiles, pottery and stones show that rubbish material was dumped here. At the bottom of stratum VI the surface of wall G1 appeared (only one block) at -670. A bronze coin (cat. no. LL-11) was found at -622. Stratum V contained a large amount of African Red Slip Ware with many forms present, the majority belonging to productions of the later 1^st to 3^rd centuries AD.

217Stratum III + IV-IVa (westernmost part) represents frequentation of the area. At the bottom of the layer the number of fragments of tile and pottery increased.

218FINDS

219G VI: many roof tiles, thin-walled ware; Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD); Italian Sigillata (Consp.9, 21, 28/29); African Red Slip Ware (Hayes Form 8A); amphorae; glass; rounded lamp (type A2a = Hellenistic), lamp (type B4 = c.50-150 AD, a perline/globetti type B5c).

220G V + IV + IVa (westernmost part): roof tiles; Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD); much African Red Slip Ware (Hayes Forms 3, 6, 8A, 8B, 9B, 23B, 27, 50?, 58, 61A); common ware casserole (type E = late 2^nd-4^th cent. AD); amphorae (unassigned); glass; lamp (Firmalampen type B3 = mid- 1^st cent. AD onwards, type B4 = c.70-225 AD, a perline/globetti type B5a = c.75 AD onwards, type B5b with short, round nozzle, a perline/globetti type B5c = c.75 AD onwards); bone (pins); coin (cat. no. LL-11: probably Antonine); metal (bronze buckle).

221G III: thin-walled ware (with rouletted decoration); African Red Slip Ware (Hayes Form 6, 61A); lamp (a perline/globetti type B5b with short, round nozzle, type B7b = c.425-650 AD); metal.

Trench H

222Excavated 1989-90, strata I-VI (Figs. 45-48)

223The trench (Quadrante 174) comprises the area 88.5-91.5/136.5-143.5 without 88.5-89.5/139.5-143.5. Stratum VI was excavated in an isolated area (90.7-91.5/136.5-140.0). The bedrock was not reached.

224Phase 1: Early Hellenistic period: Courtyard House

225Strata VI-IV (lower part) represent occupation and destruction. The two walls, HI and H2, constitute the western façade of the Courtyard House in use from the late 4^th century BC into the early 3^rd century BC. Wall HI appeared at the bottom of stratum IV (level -665/677) and wall H2 in stratum V (level -666/678). Their bottom levels were not reached. The courtyard is located north of wall H2 and is paved with large flat stones (level -661/663) (Fig. 46).

226The destruction layer associated with the building was massive, beginning at level -647/650 in stratum IV. Charcoal was found on both sides of wall H1 together with large amounts of roof tiles and pottery. Much charcoal was located in the southern part of the trench. Some of the mudbricks, east and west of wall H1, are secondarily burnt. Stucco was found above, east and west of the wall.

227Two silver coins (cat. nos. LL-1 and LL-2) were found on and immediately west of the wall, both at level -675 and point to a destruction in the early 3^rd century at the latest. A Paestan bronze coin (cat. no. LL-6) of the late 5^th century BC was found in stratum IV at -660. In this stratum the lower part belongs to the destruction layer continuing in stratum V-VI, and the upper part to a later phase of frequentation.

228FINDS

229H VI: roof tiles; stucco; black-gloss ware; core-formed glass (rim-disc from alabastron); metal (bronze).

230H V: large quantity of roof tiles, some with paint; many mudbricks; stucco; South Italian red-figure pottery, much black-gloss ware (Morel 1315, 1520, 2283, 2284, 2570/2580, 2700, 2787, 4100, 4373?, 4382); overpainted wares; red-gloss ware (Morel 2251, 2981); common ware casserole (type A = Hellenistic/Republican); coarse ware (louterion); coins (cat. nos. LL-1 and LL-2: Campanian, last quarter of 4^th to early 3^rd century BC); bronze fragments; iron (fibulae, head of axe, studs, wall-fixtures).

231H IV (finds from upper and lower part listed together): roof tiles; mudbricks; stucco; black-gloss ware (Morel 2285, 2671, 2770, 2932), black-gloss ware with overpainted decoration, brown-gloss ware; Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD); Italian Sigillata (Consp.8.1, 23); African Red Slip Ware (Hayes Form 6); common ware casserole (type A, Hellenistic/Republican), lids with top and orlo annerito; lamp (type B4 = probably 50 AD-beg. 2^nd cent. AD and c.80-150 AD, a perline/globetti type B5a); terracotta statuette; loomweight; coin (cat. no. LL-6: Paestan, late 5^th cent. BC); bronze handle; iron (chisel, nail, stud).

232Phases 3-4: End of Republican period-imperial period

233Stratum IV (upper part)-III represent a period of use/frequentation of the area.

234In stratum IV the finds span the period from the later part of the 1^st century BC into the 1^st/2^nd century AD. A diagonal channel or furrow cut through the northern part of the trench (140.5-142.5, -658/659). Human bones were found here.

235Stratum III contained collapse material, for instances much tile and pottery. Lime sediment (a kind of channel?) was located in northernmost part of the trench (c.142.5-143.5) (Fig. 47). Here many tiny glass fragments and tube-shaped lime sediments were found. The latter also occurred in a zone a little further south (c.140.5-141.5).

236A child burial ad enchytrismos (Tomba 6008) was dug into stratum III at level -629.5 (Fig. 48). The amphora is Iberian and belongs to Peacock & Williams Class 23. It was produced in the period between the 3^rd and early 5^th century AD.

237FINDS

238H III: Pompeian-Red Ware (end 1^st cent. BC-beg. of 2^nd/3^rd cent. AD); African Red Slip Ware (Hayes Forms 6, 8A, 58); common ware casserole (type E = late 2^nd-4^th cent. AD); glass; lamps (a perline/globetti type B5b, type B5c); iron key.

Trench J

239Excavated 1990, strata I-V (Figs. 49-51).

240Trench J is rectangular and covers the area within 89.5-91.5/144.5-147.0. The bedrock was not reached.

241Phase 1: Early Hellenistic: Courtyard House Strata V-IV (bottom) represent occupation and destruction. The walls J1 and J2 and an apparently stone-paved surface north of wall J2 (level -684/706) belong to the north-western part of the courtyard house (Fig. 50). The surfaces of the walls appeared at the bottom of stratum IV (tl -660 and-657 respectively). Their bottom levels were not reached. In stratum IV two layers of collapse material were located (east of wall Jl). They are clearly distinguishable from each other due to a layer of soil in between. Moreover, the fragments of roof tiles were bigger in the lowermost, and there was a layer of soil between the two layers of collapse material.

242Two roof tiles with their entire width preserved derive from stratum V immediately to the east of wall Jl, one at level -709. These are clearly part of the collapse material fallen on a surface, and a floor level (and street?) should be expected some 20.0-40.0 cm further down. In the southern part of the insula (trench G) the street is located at -738/740.

243This phase of occupation can be dated to the late 4^th to early 3^rd century BC.

244FINDS

245J V: roof tiles, some with paint; stucco; black-gloss ware; common ware jar (type 7), lid (type A6); coarse ware dolium (type F3 and unassigned).

246Phases 3-4: End of Republican period-imperial period

247The upper part of stratum IV contained collapse material, which cannot be associated with any structures. Probably the material was dumped here or it came from the building west of the street, belonging to the second phase of occupation.

248Stratum III contained collapse material (tl -589/606, south-eastern corner) of both Republican and Imperial date. It is not related to any structures and must have been brought here sometime during the Imperial period. Lime sediment was located in the southernmost part of the trench.

249A child burial ad enchytrismos (Tomba 6093) was dug into stratum III, bottom level at -637 (Fig. 51). The amphora is a Gauloise 4 of the mid-1^st to 3^rd century AD.

250FINDS

251J IV: black-gloss ware (Morel 2284, 2424); red-gloss ware; common ware jar (type 5); coarse ware basin (type B), dolium (type E and unassigned), mortarium (type B2); loomweight.

252J III: black-gloss ware (Morel 1312, 2640/2650); thinwalled ware; Pompeian-Red Ware (end of 1^st cent. BC-beg. of 2^nd/3^rd cent. AD); African Red Slip Ware (Hayes Form 8B, 23B, 50? 61A); amphora (Dressel IB = c.70-10 BC); glass; lamp (type B4 = probably last quarter of 2^nd/first quarter of 3^rd cent. AD).

Trench K

253Excavated 1989-90, strata I-VI (Figs. 52-53).

254The rectangular trench (Quadrante 174) comprises the area within 92.5-95.5/136.5-138.5. No structures in situ are located here, and the bedrock was not reached.

255Phase 1: Early Hellenistic period: Courtyard House

256Strata VI-IV (below terra battuta) represent the occupation/use of the zone until the destruction of the Courtyard House in the early 3^rd century BC at the latest. The lowest layer, stratum VI, contained a large amount of unworked blocks (level -674/688), which might come from a collapsed inner wall/structure in rubble masonry (Fig. 53). The many fragments of ridgepole tiles in stratum V testily to the location of the ridge of the roof above this zone aligned either NE/SW or SE/NW. Apparently the collapse of the wall/structure was followed by the collapse of the roof. An uneven layer of terra battuta was registered at level -657/668 in stratum IV. In the eastern half of the trench the terra battuta was dug through, revealing the massive destruction layer (tl -659/661), and both mudbricks and large quantities of ashes appeared. A bronze coin, cat. no. LL-4, was found at the bottom of the stratum (-663). It probably dates to the first half of the 3^rd century BC. The question is whether the coin forms part of the destruction layer or whether it was dropped at a later time. Stratum VI contained some residual material from the 6^th-5^th centuries BC.

257FINDS

258K VI: roof tiles; Attic black-figure pottery; black-gloss ware of some quantity and quality (Attic-type cups, Morel 4311).

259K V: massive tile fall, including almost all the ridgepole tiles found in Prop. Avallone; red-figure pottery; blackgloss ware (Morel 2285); overpainted wares; common ware casserole (type A = Hellenistic/Republican).

260K IV (below terra battuta): common ware jar (type 4d, parallels from Pompeii, mainly Hellenistic layers); coin (cat. no. LL-4: probably first half of 3^rd cent. BC).

261Phases 3-4: End of Republican period-imperial period

262Stratum IV: The terra battuta layer was made in order to seal the collapse material from the first phase of occupation. The layer consists of very hard clayish soil with a large amount of small fragments of pottery and tile. The area continued to be frequented or used (as a dump?) as shown by the later material from the upper part of stratum IV. Here the soil was compressed.

263Stratum III represents a period of frequentation in the area dated to the Imperial age. The layer had been disturbed and contained late Republican to Imperial finds.

264FINDS

265K IV (above terra battuta): lead-glazed ware (ribbed handle); African Red Slip Ware.

266K III: African Red Slip Ware; common ware platters orlo bifido (type l/5a = late 2^nd cent. BC-lst cent. AD), lid orlo annerito (type Bl); amphora (Dressel IB = c.70-10 BC and unassigned).

Trench L

267Excavated 1990, strata I-V (Figs. 54-56).

268The trench (Quadrante 174) covers the area 96.5-99.5/136.5-140.5 and the square 99.5-100.5/139.5-140.5. The bedrock was not reached.

269Phase 1: Early Hellenistic period: Courtyard House

270Strata V-IV represent the occupation and destruction of the courtyard house whose walls (L1-L3) appeared in stratum III and continued in strata IV-V (tl LI:-651/-687; tl L2:-653; tl L3:-651/654). The bottom levels of the walls were not reached, and no remains of floors or paving were revealed. The strata are characterized by a thick layer of collapse material, concentrated in the area to the east of wall LI and south of wall L3. Here a large dolium (cat. no. AA-29) and 16 loomweights (13 from stratum IV and 3 from stratum V) were found. The dolium could be almost wholly reconstructed and must have been destroyed when the building collapsed. No traces of fire were found on the objects.

271Judging by the datable finds, the occupation spans a period from the late 4^th into the 3^rd century BC. It is also clear that there are later disturbances in stratum IV, e.g. postholes and isolated zones with soil of a different colour. The finds from the upper part of stratum IV and these zones date to the Imperial period.

272FINDS

273L V: black-gloss ware; coarse ware basin (type B); loomweights.

274L IV: roof tiles; mudbricks; stucco (polychrome); South Italian red-figured pottery; black-gloss ware (Morel 1520, 1552, 4372?, 4384), black-gloss ware with overpainted decoration; Pompeian-Red Ware (end 1^st cent. BC-beg. 2^nd/3^rd cent. AD); African Red Slip Ware (Hayes Form 9B); common ware casserole (type A = Hellenistic/Republican); coarse ware dolium (types D and I, and unassigned), situla; amphora (unassigned); 13 loomweights; terracotta disc = theatre ticket?; silver coin (cat. no. LL-3: Campanian, late 4^th or early 3^rd cent. BC); bronze vessel; iron nails.

275Phase 4: Imperial period

276Strata IV (upper part)-III are interpreted as a disturbance of the area with different types of soil and ‘postholes’ sometime during the Imperial period. In the south-eastern comer collapse material (roof tiles and stones) probably forms part of the destruction layer found in trench M.

277The surfaces of the walls appeared at the bottom of stratum III.

278A child burial ad enchytrismos was dug into stratum III (Tomba 6092, level -627/637) (Fig. 56). The amphora is an Africana I/Piccola, produced from the second half of the 2^nd to 4^th century AD.

279FINDS

280L III: some quantity of thin-walled ware; African Red Slip Ware (Hayes Forms 8B, 9B, 23B, and unassigned); amphora (Dressel 20 = Tiberian/Claudian- 3^rd/4^th cent. AD and unassigned); cast glass; lamp (a perline/globetti type B5d); terracotta statuette.

281General stratigraphy for trenches H, J, K and L: Courtyard House (Fig. 57)

282The stratigraphy is coherent in the four trenches with strata VI-IV (lower part) representing the earliest occupation of the area, manifested in the Courtyard House, and followed by a destruction level. The western façade of the house is documented in the coordinates from 136.5 to 143.5 and from 144.5 to 147.5 (walls H1-Jl), i.e. a stretch of 10.0 m or 11.0 m including the unexcavated balk between trenches H and J. 3.0 m of the eastern NE/SW wall (probably not a facade) was revealed in trench L (wall LI). Moreover, transverse walls, i.e. aligned NW/SE, testily to internal partitions (walls H2-L2, J2, L3). The extension towards the north is not clear, nor is it certain whether the zone east of wall LI was roofed or unroofed, but the amount of roof tiles speaks in favour of the former. Towards the west the building faced onto a street (not attested). A large block was found in this supposed street zone (trench J).

283The surface of the western wall (walls Hl-Jl) was reached at the following levels at the bottom of stratum IV: trench H-665/676, trench J-660; the surface of the eastern NE/SW wall (wall L1) was reached at the bottom of stratum III at -651 and at the bottom of stratum IV at level -668. Surface levels for wall H2-L2 are-666/678 in H V and-653 in L III (shift to stratum IV); wall J2-657 in J IV and wall L3-651/654 in L III.

284The only floor level was reached in H V north of wall H2 at -661/665 where a courtyard has been identified. Large flat stones form the paving. Since no other floors were found, it is likely that the courtyard is located on a higher level than the rest of the rooms belonging to the house.

285North of wall J2 another paved area is likely to have existed. Here flat stones and large pieces of roof tiles were found (-684/706).

286Concentrations of collapse material, including massive tile falls, began at c.-647/650 (trench H), -635/656 and-646/659 (trench J), -659/661 (trench K) and -656 (trench L). This layer is generally characterized by hard, sometimes clayish, soil. Traces of fire from the destruction were found on objects in trench K (including also ashes and secondarily burnt mudbricks) and H, whereas trench L revealed no such.

287The finds from the destruction layer date the time of occupation to the second half/late 4^th to early 3^rd century BC. Foundation trenches were not found/reached. The lowermost part of K VI contained material of the late Archaic to Classical periods. There is also a Paestan coin of the5^th century BC from H V.

288Frequentation and burials

289Strata IV (upper part)-III: After the destruction of the Courtyard House the area in trench K was sealed with a layer of terra battuta. The later destruction in the first half of the 1^st century BC might be represented by the upper destruction layer in J IV.

290The frequentation and disturbance of the area in the Imperial period is evidenced by the material from the upper part of stratum IV in all the trenches (e.g. Pompeian-Red Ware platters, lids orlo annerito, lamp types B4 and B5a), and in stratum III rubbish material was present in trenches H (-579/590), J (-589/606) and L (-627/642). Most likely the zone was used as a kind of dump.

291Three child burials ad enchytrismos were dug into stratum III. They were situated at the same level (-629.5 in trench H, -625/637 in trench J and-624/637 in trench L) and testify to the use of the area as a burial ground.

Trench M

292Excavated 1986, strata I-IV (Figs. 58-59).

293The trench (Quadrante 174) covers the area 100.5-106.5/128.5-130.5 and 100.5-103.5/130.5-131.5. The bedrock was not reached.

294Phase 4: Imperial period

295Stratum IV is defined by the two walls: Ml (orientated SE/NW, -606) and M2 (orientated NE/SW, -618). Their bottom levels were not reached, and the time of construction cannot be dated. Much charcoal was found between the blocks and next to them. The layer was rich in finds, particularly in glass, but generally it is very mixed with finds ranging in date from the later 1^st century to the 3^rd and 4^th centuries AD. It is thus possible that the structure goes back to the later part of the 1^st century AD, but whether the material represents occupation or frequentation is hard to determine.

296Stratum III represents a period of frequentation. The material mainly belongs to the 3^rd and 4^th centuries AD. A few medieval and modem sherds were located in the upper part of the layer.

297FINDS

298M IV: some terra sigillata sherds (Consp.9 and unassigned); African Red Slip Ware (Hayes Form 8B, 50?); common ware casserole (type E = late 2^nd to 4^th cent. AD); glass; lamps (type B4 with short round, nozzle = mid-1^st to 3^rd cent. AD, a perline/globetti type B5c, type B7a = African); coin (cat. no. LL-14: Theodosius I, 379-383 AD).

299M III: African Red Slip Ware (Hayes Form 3A); common ware casserole (type E = late 2^nd-4^th cent. AD); glass; lamp (types B3-B7b).

Trench N

300Excavated 1986 and 1989, strata I-IX (Fig. 60).

301This square trench (Quadrante 160) comprises the area within the coordinates 120.5-123.5/156.5-159.5. The stratigraphy and character of trench N are quite different from the rest. Apart from a double row of thin stone slabs in stratum III, no structures were revealed. Strata II-III had the highest contents of finds. In the following layers the number of finds decreased and ceased completely in the lowest levels, strata VIII (lower part)-IX.

302Stratum IX was excavated in an isolated area (120.5-122.0/157.5-159.5).

303The extreme fragmentation of the finds in strata II-III shows that the material was brought to the area, and at least as far as the Orientalizing and Archaic material (pottery) is concerned, came from destroyed tombs, as the repertoire of shapes matches the one found in graves. The strata, however, contained a complete mix of material from the early Iron Age or Archaic period to the late Imperial period; most probably strata II-III represent a levelling of the area, which may have taken place in modem time as well as in ancient. The few finds of early material in strata IV-VII should probably be explained by the natural displacement of artefacts (e.g. caused by animals).

304Stratum IX: All over the excavated area the yellow travertine bedrock was reached. It was full of hollows and grooves. The layer was void of finds.

305Stratum VIII: The soil was greasy and clayish with white lime flecks and black charcoal flecks. A line of tiny white stones appeared in the north-eastern comer at -692. In the south-western corner clayish spots were located at -703/708/712. The lower part of the stratum was void of finds.

306Strata VII-V: These layers contained finds dating from the Archaic period (bucchero, argilla depurata) to the Hellenistic/Imperial period.

307Strata IV: The layer is a continuation of stratum III. The content of pottery was low and decreased in the lower part of the stratum. Finds date from the early Iron Age (impasto) to the Hellenistic period.

308Stratum III: The surface of collapse material consisting of roof tiles, pottery and stones appeared at c.-625/626. This collapse material was located in isolated areas in the northern and southern part of the trench. The layer also contained a wall/tomb structure made of thin stone slabs, which were placed in a double row and put on the edge, inclining nine degrees towards the east. The slabs were almost on a N/S alignment (bottom level at -635). The slabs are 17.5 cm high and c.2.3 cm thick.

309FINDS: impasto, bucchero, argilla depurata, black-gloss ware, black-gloss ware with overpainted decoration, thinwalled ware, lamp (type B4), spindle whorl.

310Stratum II: The layer has been disturbed in modem times.

311ANCIENT FINDS: impasto, bucchero, argilla depurata, Corinthian, black-gloss ware, black-gloss ware with overpainted decoration, thin-walled ware, terra sigillata, African Red Slip Ware (unassigned fragments), coarse ware dolium fragments, Ionic-Massiliotic amphora, free-blown glass, iron nails.

312Stratum I: The stratum has been disturbed in modern time, containing both ancient (from Hellenistic time onwards) and modem finds.

II.2 Architectural remains: walls

313Birgit Tang

314Stretches of walls forming the socle courses of the buildings were found in all trenches except for trenches K and N (Figs. 4 and 61-63). However, the tile fall located in trench K identifies the excavated area as a roofed space. The walls are aligned NE/SW and NW/SE. For convenience of description, e.g. when relating the walls to each other, a conventional north is used here, corresponding to north-east.

315No structure or building was fully excavated, but the alignments of the walls suggest that there are at least two blocks or insulae, one to the east and one to the west of a c.6.5 m wide street paved with terra battuta and aligned NE/SW. Moreover, the walls situated in the north-eastern zone, in trenches H, J, K and L, belong to a courtyard house.

316Phase 1: Layout of the quarter and Courtyard House (Fig. 61)

317The earliest sign of building activity is the terra battuta street at -785/788 in G XIV+XVI. This street was cut when the eastern insula was constructed, i.e. when the general layout took place. The terra battuta street belonging to this phase is located at -738/740 (G XII) and at -748 (Ds X/XI).

318Walls belonging to the first phase of occupation are located both east and west of the street, but they are best preserved east of the street: trenches A?, E, G, H, J and L. In trenches D and F the lowest part of the walls probably goes back to this phase as well, but due to reuse in the succeeding phase, their first phase is not clearly distinguishable (cf. below).

319Judging from the foundation deposit in trench G, the quarter was laid out in the second half of the 4^th century BC. The time of destruction is dated by the collapse material from trenches G-L; namely at the latest in the early 3^rd century BC. Moreover, some finds from trenches A, D, E and F seem to relate to this first occupation phase, thus indicating the existence of the buildings in this period.

320The buildings were provided with substantial socle courses in ashlar masonry. The ashlar blocks may be over 1.0 m in length. This applies to the transverse walls as well. Another characteristic feature is a semicircular or circular cutting in the blocks. Here timber posts were placed.

321The only floor level reached in the northern part of the excavation area is the paved courtyard in the Courtyard House (H V) at -661/663, but the find of two roof tiles at level -709 in J V, immediately to the east of wall J1, indicates that floor paving is situated below this level; especially as large flat stones were registered at -684/706 in the same area. They may have formed a paving, which would have been on a slightly lower level than the one in trench H.

322In the eastern insula the built-up area is documented by the NE/SW façade wall (wall Gl) southernmost in trench G at 126.0 (tl -669/675) and northernmost in trench J at c. 147.0 (wall Jl, tl -660) with nonexcavated areas from 128.0 to 136.5 and from 143.5 to 144.5. The bottom level for wall Gl is -795. There is a minor divergence in alignment; the southern stretch lying 89.5-89.95 and the northern 90.2/90.24/90.28-90.68/90.72/90.80. The extension of the eastern insula towards the east cannot be determined with certainty, but in trench L the area to the east of the NE/SW wall (wall LI) most likely represents a covered space. The Courtyard House located in the northern part of the eastern insula (trenches H, J, K and L) would then be more than 7.20 m wide (walls included) and has a courtyard measuring c.6.20 x 4.50 m. Ridgepole tiles from trench K show the location of the ridge of the roof (Fig. 64).

323To the west of the street the first phase of occupation is documented by the NE/SW wall (wall Dl) recovered in trench Ds (123.0-125.0, c.-700) and in trench E (wall El) (129.7-131.38, -760/772) constituting the façade of the western insula towards the east. Due to the reuse of the walls for later structures, this first phase is poorly preserved and harder to detect. Again the alignment of the walls differs, ranging between c.82.7-83.22 in trench Ds and 83.1-83.5 in trench E. The bottom level for the wall in trench E, -793/799, corresponds with the one in trench G (-795), while the one in trench Ds is at -748. Wall A1 also seems to go back to the first phase. Its surface level is at -713/735.5. Whether the wall formed the southern façade of the insula cannot be answered.

324Description of the structures: western insula

325Trench A, stratum VII: The double wall aligned NW/SE (wall Al, -713/735.5) and a burnt clay floor (-734) are the only remains from this phase (Figs. 5-6). The excavated part of the wall is c.2.15 m long (77.5-c.79.65/c.118.6-119.5); its width is c.80.0 cm. The middle block in the southern section of the wall exceeds one metre in length (1.12 m) and has a semicircular cutting for a post. A second posthole is located at the southern balk (-748) almost aligned with the first one. The burnt floor is situated in a gap of the northern section of the wall, which might be identified as a door opening. Apparently, the travertine bedrock, which is visible at various points south of the wall, was incorporated into the wall, since there are no proper division of the blocks.

326Trench Ds, stratum XI: The lowest course of the NE/SW wall (wall Dl) built in limestone belongs to the initial phase to some extent, as well as the lowest terra battuta layer of the street.

327Trench E, strata X-IX: The stretch of the NE/SW wall (wall El) located in c.129.70 (the southern block is only partly visible) to 131.38 at level -760/772 seems to be the continuation of wall Dl in trench Ds, although the bottom levels of the two vary considerably. This, however, can be explained by the highly uneven surface of the bedrock. Three large blocks constitute the stretch, their width being c.40.0 cm (83.1-83.5) and height c.30.0-35.0 cm (Figs. 22, 25 and 32). The excavated length of the northernmost block is 76.0 cm, and the central block c.40.0 cm (Figs. 27 and 29). The bottom level for the two northernmost blocks is -793/799. It is not clear whether the wall rests on the bedrock or not.

328Trench F, stratum XI: The NE/SW wall (wall FI) documented in 83.4-83.9/136.5-139.5 most likely contains blocks from the first phase or a larger part of the wall might be original. Three large blocks in its southern half can be paralleled with the blocks used in the other trenches (Figs. 34 and 36). The northernmost seems to have been placed upright as part of the rebuilding (in this position: 70.0 cm high and c.20.0-32.0 cm wide). The two other blocks are 44.0 cm high.

329Description of the structures: eastern insula

330Trench G, strata XV-VII: Here the NE/SW wall (wall Gl) in travertine is documented in 126.0-128.0 with top level at -669/675 (stratum VII). Only the outer side, i.e. the western side, was excavated (Figs. 24-25). Two large blocks c.40.0-45.0 cm wide form the uppermost course (89.5-89.95). The southern block measures 1.21 m in length. Between 126.0 and 127.0 the wall was excavated in its entire height, c.1.20-1.25 m (bl -795). The northern block has two drilled holes near its northern end, c.9.5 cm and 8.0 cm deep respectively. Moreover, there is a rectangular recess, c. 10.0 x 5.0 x 1.0 cm, at its southern end. Except for the smoothed southern block, the surfaces are very corroded (Fig. 42). A drainage hole/outlet is situated in the wall.

331The street in stratum XII belonging to this phase is documented to the west of the wall at -738/740.

332Trench H, strata VI-IV (Courtyard House): The excavated part of the NE/SW wall (wall HI) comprises 7.0 m (136.5-143.5) with tl at -665.5/667 and-676/678 (Fig. 45). As usual the width is c.40.0-45.0 cm, lying between coordinates 90.20-90.72, and the length of some of the blocks exceeds 1.0 m. One of these has an almost circular cutting on the outer side. A wooden post was placed here in order to support the upper part of the wall built in mudbrick. Only c.80.0 cm of the inner transverse wall aligned NW/SE (wall H2) was excavated, tl at -666/678, width c.40.0-45.0 cm (c.90.70-91.5/140.0-140.46). It joins wall HI by way of cutting. The bottom level for the walls was not reached. Both are built in travertine.

333Immediately to the north of wall H2 an open area can be identified judging from the pavement, which consists of large flat stones located at level -661/663.

334The function of the large block located to the west of wall H1, i.e. in the supposed street, is unclear.

335Trench J, strata V-IV (Courtyard House): The NE/SW wall (wall Jl) is documented for another 3.0 m, 144.5-147.5, i.e. the whole length of the trench (Figs. 49-50) at 660 (90.24/90.28-90.68/90.80). However, the two small stones at the northern balk do not seem to be part of the wall. Again one of the large blocks is provided with an almost circular cutting (the southern). This block, the block north of it and the two small stones are clearly not in their original position, which may indicate that the wall ends here. The two large blocks are 40.0 cm wide. The transverse wall aligned NW/SE (wall J2), lying between 90.75-91.5/145.0-145.4, level -657, is 40.0 cm wide and with an excavated length of c.75.0 cm. North of this wall a second stone-paved area might be located. Large flat stones were found together with large fragments of roof tiles. Walls H2 and J2 create a space c.4.50 m long. The bottom level for the walls was not reached.

336Trench K, strata VI-IV (Courtyard House): As mentioned earlier the trench is void of structures, but the many roof tiles, especially ridge pole tiles, found here show that it was a covered zone (Figs. 52-53).

337Trench L, strata V-III (Courtyard House): Together with the western NE/SW wall (wall HI) the NE/SW wall (wall L1, -651/668) here forms a covered area c.6.20 m wide (within the walls) (Figs. 54 and 57). Whether this wall is also to be identified as the eastern limit of the insula is uncertain, but the amount of collapse material, e.g. fragments of stucco, located to the east of the wall speaks in favour of a continuation of the building. The excavated length is c.3.80 m (136.7-140.5), and again the characteristic roughly semicircular cutting is present, this time in the southernmost block (60.0 cm wide). The largest block has a width of c.68 cm, while the central blocks are smaller with a width of c. 40.0-50.0 cm (varying within the coordinates 96.84-97.04-97.44-97.56). A transverse wall (wall L2) is documented west of wall L1 beginning at -653 and excavated to the length of c.35.0-40.0 cm (96.5-c. 96.85/139.72-140.36). It is situated a little further to south than the transverse wall in trench H (wall H2). The large block (wall L3) to the east of wall L1 is 1.34 m long (-651/654) and can be interpreted as a fallen block or as part of a wall. Its alignment with the western transverse wall speaks in favour of the latter. The bottom level for the walls was not reached.

338Phase 2: Rebuilding in the western insula (Fig. 62)

339This phase is documented in the insula west of the street only where the earlier structures were reused. The structures are located in trenches B, C, D, E and F. In trench E new courses were constructed on top of the older walls, whereas the earlier walls in trenches Ds and F were incorporated to a larger extent in the new ones. The street is documented in trenches Ds (XI/X-748, renewed) and E (VIII between 741/752 and-751/756).

340Generally, the walls are built of limestone blocks, but reused material and roof tiles are also used as building material, e.g. wall Bl (blocks with stucco in situ) and in walls D1-D2 (basin). Although there are only a few metres between the trenches, it is not possible, apart from the outer walls facing the street, to connect the various remains of walls, nor to conclude anything about room distribution.

341The reoccupation seems to have taken place sometime during the 2^nd century BC. The finds from the collapse material date the time of destruction before the mid-1^st century BC.

342The eastern façade is documented from 122.0127.5 (trench Dn + Ds, -675/686) and from 136.5139.5 (trench F, -671) (Figs. 10-12, 34 and 35). In trench E the façade (wall E2) was demolished, but the clearly linear eastern limit for the tile pavement (at 83.05), the stucco fragments and the dressed block partly beneath the pavement indicate the presence of a wall (Figs. 22 and 27). Wall F1 is located a little further to the east than wall D1, 83.4-83.9 compared with 82.7-83.28.

343Internal partition of the insula is documented in trench Ds where a transverse wall (wall D2) was constructed (-714/719), and a terra battuta floor was laid (-731/751) north of this wall. The L-shaped wall (wall E4) and tile pavement were made in trench E (-710.5/726). There is also a transverse wall in trench F (wall F2) (-688/703).

344The western limit of the insula was not located. The westernmost walls are walls B1 and Cl at 76.76, i.e. more than 6 m from the eastern façade wall. Clayish layers or floors are present in trench B at -733/738 and-750/752. Trench B is situated immediately to the west of trench E.

345Description of the structures:

346Trench B, strata VII-IV: Remains of two NE/SW walls were found, one at the western balk (wall B1, -673/678) and one further to the east (wall B2, -711/714) (Fig. 7). The excavated length of wall B1 is c. 1.3 m and the width c.40.0 cm; it is c.68.0 cm high (76.75-77.15/130.08-131.38). The wall contains reused material: a squarish block with stucco in situ. Two blocks at the northern balk constitute the eastern wall B2. They are c.40.0 cm wide, and the excavated length is c.64.0 cm (78.10-78.52/130.72-131.36). Terra battuta or clay floors are located to the east and south of wall B2 at level -733/738, and a second one between walls B1 and B2 at -750/752 (Fig. 8). Wall B1 is directly overlying the clay floor.

347Trench C, strata VI-V: It is not clear whether the remains of a NE/SW wall in the northern and southern (wall C2, -698/701 and-677/682) part of the trench belong together (Figs. 7 and 9). The southern one has an excavated length of c.90.0 cm, and the width is c.35.0 cm (78.14-78.5/132.56-133.46). The northern one is a looser structure with blocks spread in an area c.1.4 m long (c.78.0-78.5/133.98134.4). These walls are more or less on alignment with wall B2. Moreover, a NW/SE wall (wall Cl, -670/700) can be identified in the southern part, apparently meeting the southern part of wall C2 at a right angle and thereby forming a comer (blocks in the area 76.75-78.1/133.0-133.58).

348In the northern part of the trench a large worked limestone block appeared within the coordinates 77.5-78.2/134.0-135.5 at level -688/693. Moreover, there is a kind of channel at level -697/704 immediately west of the worked block.

349Trench Dn, stratum VI + Ds, strata XI/X-IX: The excavated stretch of the NE/SW wall (wall Dl) is c.5.5 m with top levels at -676.5/686 (Dn) and-675/677 (Ds), and a width of c.35.0-55.0 cm (82.70-83.06 (south)/82.74-83.28 (north)/122.0-127.5) (Figs. 10-12). Thus the total height is c.73 cm (bl -748). In Dn excavation stopped c. 10.0 cm below the surface of the wall. In Ds the earlier wall was repaired with small stones and rubble, whereas the uppermost course is built of large relatively flat stones with a height of c. 15.0-20.0 cm (Fig. 13). The lowest course is located between c. 123.0-125.0, i.e. it is only documented in the southern part of the trench at c.-700. The blocks measure from 20.0-24.0 cm in length and are 15.0-24.0 cm high (c.82.76-83.10 [at 123.0] and 82.70-83.22 [at 125.0]). The height of the course is c.50.0 cm with bottom level at -748. The wall is provided with an outlet, which is triangular on the inner side (here it seems to be partly blocked?) and more or less rectangular on the outer side (Figs. 17-18). This drain seems to be connected with the pit dug in the bedrock in the street, measuring c.30.0 cm in depth (Figs. 15 and 19). Most likely these features were made in this second phase of occupation or they were repaired. This assumption is strengthened by the fact that the lowermost course is partly rebuilt/altered, as it contains a basin used as comer stone for the transverse wall (D2).

350The inner transverse wall (wall D2, tl -714/719 and bl -755) to the west of wall Dl is partly constructed of reused material with an excavated length of c.1.70 cm and a width of c.45.0-50.0 cm (c.81.0-82.7/123.5-124.0). A small earth fill is located between the bedrock and the wall. The outlet in wall D1 was either constmcted or it was repaired (cf. above).

351Terra battuta paving is documented in the street (upper layer, -748) and inside the house (-731/751).

352Trench E, strata VIII-V: A new NE/SW wall (wall E2) was built facing the street towards the east. Only one ashlar block is preserved with top level at -726 (Fig. 27). It is c.30.0 cm high and with an excavated length of c.55.0 cm and width of c.40.0 cm (83.16-83.56/129.7-130.25). The wall follows the alignment of the wall in trench D. Inside the building an L-shaped wall (wall E4) and a tile pavement (-710.5/726) were constmeted (Fig. 31). The wall is c.40.0-45.0 cm wide and is composed of small blocks laid in more courses in the northern stretch (c.1.9 m long), while the western stretch with a preserved length of c. 1.3 m combines large blocks and rubble (Figs. 28-29) (80.78-81.2/130.04-131.34 (western); 80.9-82.8/131.0-131.40 (northern)). The pavement, composed of reused roof tiles, is situated in the area 81.25-83.05/129.6-131.0. The room was provided with a drainage system consisting of a cover tile on the floor and an outlet in the outer wall (only visible in the balk, bottom level -754) (Fig. 33).

353The terra battuta remains located between -741/752 and-751/756 east of wall E2 are likely to be identified as a street.

354Trench F, strata XI/X-V: After the first destruction the area was cleaned up, and wall F1 was rebuilt/repaired. It is c.50.0 cm wide with highest surface level at -671, and was excavated from 136.5 to 139.5 (83.4-83.9) (Fig. 34). The wall is not completely aligned with the wall stretches in trenches D and E. Both large and small blocks were used in the construction; the lowest course in its southern part is built in ashlar masonry (e.g. three large blocks, probably reused from the first phase of occupation), whereas smaller unworked limestone blocks were employed for the upper course and its northern part (Fig. 36). This reveals the two phases of construction. The posthole situated in the wall probably supported the upper part of the wall built in perishable material. The second wall is a transverse wall (wall F2) (-688/703, excavated length c.2.90 m, not excavated in the central part, c.45.0-50.0 cm wide) meeting wall F1 in a right angle on its western side and built in unworked limestone blocks, also using small stones and fragments of tiles (80.5-83.4/137.96-138.04-138.5) (Fig. 37). Wall F1 is lying directly on the bedrock, bottom levels at -720/728 for the northern part and-748/754 for its southern part. The lowest course of the transverse wall F2 is difficult to detect (only excavated to the level of the bedrock between 82.5-83.4), but it is lying on soil and the lowest larger block is at c.-732.

355Phase 3: Renewal of street

356A new layer of terra battuta was laid in the street (E V-679/680, F V-679, G VIII-688/695). Additional alterations took place in trench E where wall E2 east of the tile pavement was pulled down, and a new one (wall E3) was built further to the east (tl -670/679) (Figs. 22 and 25). Probably wall E3 functioned as a kerb, as it is more like a stone edging than an actual wall.

357Description of the structures:

358Trench E, stratum V: Wall E2 was demolished; only one dressed block is left. Wall E3 was built further to the east, c.90.0 cm (tl -670/679, bl -689). The excavated length is c. 1.65 m and the width c.50.0-55.0 cm (84.0-84.52 (south)/84.1-84.64 (north)/129.70-131.5). There is no proper foundation for this wall (Fig. 26). Terra battuta is documented in the easternmost part of the trench at -679/680 (c.86.5-87.5/129.5-131.5) (Figs. 24-25).

359Trench F, stratum V: The terra battuta street is registered in the easternmost part of the trench at -679 (c.85.1-86.5/137.5-139.5) (Fig. 35).

360Trench G, stratum VIII: The terra battuta layer at -678/695 sealed the destruction layer belonging to the first phase of occupation (Figs. 24-25 and 41). Stratum VIII above the terra battuta is part of stratum VII. Here a coin (cat. no. LL-8) dated to the late Republican period was found at c.-670 northeast of wall Gl.

361Phase 4: Walls, frequentation

362Building activity was very limited in this phase and comprises a few walls in trenches A and M: wall A2 (upper courses only?), wall A3, walls Ml and M2. The exact time of construction cannot be dated, and it is possible that wall A2, or a part of it, goes back to Phase 3. The finds, however, mainly belong to the mid-and late Imperial periods.

363Description of the structures:

364Trench A, strata V-III: Remains of two walls are located in the eastern part of the trench (walls A2 and A3) (Fig. 5). The excavated length of wall A2 is c.3.0 m (tl -571/630) with a gap at the centre and a width of c.40.0-45.0 cm (within coordinates 79.68-80.08-80.16/116.5-119.5). Stucco was found in situ on its western side at level -674 (79.6/118.8). East of this wall a (curved?) NW/SE wall (wall A3) was excavated to the length of c.80.0 cm (tl -588/606), 80.08-80.88/117.22-117.36-117.74. Both walls are built of unworked limestone blocks measuring c. 10.0-40.0 cm, but their interrelation is not clear. Wall A2 is attested in strata V-III and wall A3 in stratum III.

365Trench M, stratum IV: The preserved length of the NW/SE wall (wall Ml) is c.3.6 m (tl -606) with a width of c.50.0-60.0 cm (Fig. 58) (101.64-105.24/129.72-130.32). It is not clear whether the wall continued further to the east and/or the west, but it is possible that it formed a comer with the NE/SW wall (wall M2) (tl -618) in the western part of the trench. This wall is poorly preserved, occupying an area c.70.0 cm long and c.25.0 cm wide (101.52-101.76/128.5-129.2). The excavations stopped at -624/632 before reaching the bottom levels of the walls. The walls are built of dressed blocks of travertine.

366The finds from the upper strata in the excavated area testify to the frequentation of the area throughout the Imperial period, but apart from the structures in trenches A and M, they give the impression that the area served as a dump.

II.3 Architectural fragments

II.3.1 Interior architectural decoration

367Birgit Tang

368In general, the interior architectural decoration, i.e. wall decoration and pavements, is poorly preserved, and only a few mural fragments and remains of pavements were found in situ. The material, however, shows variation within the use of colour on the walls and within pavement types. As regards wall decoration, the finds belonging to the first phase of occupation are extremely important, since they testify to polychrome walls. Most of the pavements types documented can be characterized as utilitarian, e.g. simple terra battuta and clay floors together with a floor made of tile fragments. Prestigious materials such as marble and other stones were also found, but only as isolated finds out of context. They consist of a few tesserae and cut pieces for wall and/or pavement revetment (opus sectile).

Wall decoration: Stucco

369Numerous fragments of stucco were found during the excavations, but only two, both white, were found in situ. They derive from the west side of wall A2, stratum V, level -674. The largest concentrations of fragments came from the collapse material inside the buildings in Ds X (west of wall Dl, level -696/730) and H V (east of wall H1, level -660/678).

370Unfortunately, the fragments are tiny and do not allow any reconstruction. Most fragments seem to be white, red and brown. The preservation of the pigments varies. On several fragments the colours have variably decayed, or faded, and many of the originally red fragments are now pink.

371The distribution according to colour is presented in Table 4. There is a clear difference between the fragments found to the west of the street and those found to the east: the former documenting only monochrome white walls and bicolour walls in white and red, the latter also polychrome walls. H V, i.e. the western part of the Courtyard House, not only yielded the largest amount of fragments, but also the greatest variety of colours: red, white, grey, brown, black and yellow (P1. II: cat. nos. A-2 and A-4). Furthermore, motifs are present.

372Design and motifs

373Due to the scantiness of the remains, it is not possible to reconstruct the arrangement of the differently coloured zones and the motifs, but there is no doubt that a decorative system imitating architecture was employed. This can be concluded from two white fragments with an incised line (P1. II: cat. no. A-l). They came from C VIB and Ds X, and may be part of panels with incised joints and margins. Moreover, two red fragments from trench H have a concave part, which could be interpreted as the beginning of a moulding. Whether the Prop. Avallone fragments belong to the ‘First Style’ (traditionally defined as wall painting imitating marble architecture) or the ‘Zone Style’ (imitating or reflecting mudbrick constructions) cannot be answered3. However, the combination of red and white points to the well-known and simple system composed of a red socle and a white main zone, which sometimes has a red border at the top.

374Bicolour fragments came from trenches H (red and brown, black and red) and L (a black and white zone adjoining in a straight line) (P1. II: cat. no. A-6).

375Leaving the Vesuvian cities aside, wall decoration consisting of white and coloured stucco together with murals in the First-Style tradition is widely documented on the Italian peninsula. In Southern Italy, as well as in other parts on the Italian peninsula, Sicily and Carthage, the earliest examples of First-Style murals date to the late 4^th century to the first half of the 3^rd century BC4. One of these is the habitation at S. Stefano at Buccino, which was provided with stucco decoration including architectural elements in its first phase (late 4^th to early 3^rd century BC)5. Later on, walls decorated in the First-Style tradition are documented in the Roman villas of Buccino, e.g. in room 1 (the main living room/tablinum) and the atrium area in the villa at Vittimose (phase 2, i.e. Augustan age onwards)6.

376Bicolour walls in red and white are known from several habitation contexts, e.g. room A7/A8 in Complex A at Roccagloriosa, phase IIB (last quarter of the 4^th century BC), and rooms 11-13 (principal rooms) of the habitation at Moltone di Tolve (second phase of the building dated in the 3^rd century BC)7. In addition, this decorative system with variations is also commonly found in Paestan tombs8.

377A fragment with brown and red has a partly preserved white motif (P1. II: cat. no. A-3). The most interesting is a fragment with a flower or more likely a leaf, which may be part of a festoon (P1.11: cat. no. A-5). The best parallels are found in the painted tombs from Pontecagnano and Poseidonia/Paestum decorated with bands, stripes, festoons and other plant motifs:9 Pontecagnano: e.g. Tomb 4453 dated to the late 4^th to the early 3^rd century BC10; Poseidonia/Paestum: Tomb 47 - Andriuolo (leaves) dated to 350 BC; Tomb 53-Andriuolo (leaves) dated to 350-340 BC; a chamber tomb from the Spinazzo necropolis (festoon made of leaves) and the Tomba del Magistrato from the same necropolis dated to the late 4^th or the 3^rd century BC11.

378CATALOGUE

379Cat. no. A-l: #89-319: Trench Ds. Stratum X. Level -710. PL II.

380c.50 white fragments. One fragment with incised line. 2.4 x 2.3. 2 red fragments.

381Cat. no. A-2: #89-269: Trench H. Stratum V. PI. II. Fragments: Red, brown, grey, white. 2 bicolour fragments (red and brown).

382Cat. no. A-3: #89-869: Trench H.

383Stratum V. Destruction layer above the wall. PI. II.

384Fragments: Red and white. Largest fragment: 4.4 x 2.4.

385Fragment with motif in red, brown and white. 2.5 x 1.7.

386Cat. no. A-4: #90-131: Trench H. Stratum V. 89.5/140.5 (c.91.2-91.5/139.75-139.90). Level -663. PI. II.

387Fragments: White, black/greyish, red and brown.

388One red fragment with part of profile. 1.7 x 1.5.

389Cat. no. A-5: #90-293: Trench H. Stratum V. East of the wall. PI. II.

390Fragments: Red, white, black, yellow and grey. At least 7 fragments are bicolour (black and red).

391Fragment with motif: Flower/leaf.

392Fragment: Red and part of profile.

393Cat. no. A-6: #90-741: Trench L. Stratum IV. Destruction layer. 97.3-99.5/138.5-139.7. PI. II.

394Fragments: Red, white and black. Fragment: Bicolour (black and white). 1.7 x 1.2.

395Pavements

396The pavements are divided into two groups according to find context, i.e. pavements in situ and not in situ.

Pavements in situ

397Stone

3981. Trench H, stratum V, east of wall H1 and north of wall H2:

399Large flat limestone blocks form the paving of an open area interpreted as a courtyard (Fig. 46). The surface level is at -661/663.

400Parallels for open areas paved with large stone blocks are found at e.g. Roccagloriosa, where there are at least four examples including the courtyard in Complex A on the Central Plateau. The same type of paving is also employed in the sanctuaries at Serra Lustrante (near Armento) and at Macchia di Rossano/Rossano di Vaglio. All examples date to the 4^th-3^rd centuries BC12.

4012. Trench J, stratum V, east of wall J1 and north of wall J2: Flat stones lying in an almost horizontal position were located below the collapse material, (level -684/706) (Fig. 50).

402Tile

4031. Trench E, stratum V, western part of trench within the coordinates c.81.25-83.05/129.6-131.0:

404The surface is composed of reused roof tiles set on edge at random with a fill of earth in between (Figs. 23, 25 and 31). The floor covers an area of c.1.8 x 1.4 m. The surface is highly uneven, -710/726, which is partly due to the sunken central part.

405This type of floor is commonly employed in service rooms and other areas exposed to water, e.g. kitchens, latrines, workrooms/workshops and corridors. In Southern Italy floors made of reused fragments of roof tiles are documented at various sites. The tiles are often laid in parallel rows or in the characteristic opus spicatum pattern (herringbone pattern)13. This applies to a tiled path in the sanctuary at Serra Lustrante (Armento) (phase I, second half of the 4^th century BC); and to the courtyard in a house at Oppido Lucano (end 4^th-beginning of the 3^rd century BC)14. Opus spicatum is also recorded in the Italian excavations in Prop. Truono from period II. 1 (end 4^th-beginning of 3^rd century BC)15. Tiles laid in a parallel pattern were used for e.g. the extension of the tiled path at Serra Lustrante (phase II, end 4^th-beginning of 3^rd century BC) and for the courtyard in a house at Laos where there is also an irregular part, probably a patching ( 4^th-3rd centuries BC)16. Tile paving is also widely employed in the late Republican and Imperial periods, in e.g. the villas at Vagni and San Mauro at Buccino and in houses at Cosa (the House of the Skeleton, first third of the 1^st century BC and the House of the Birds, Augustan age)17. Floors made of irregularly laid fragments of tiles are also known from Delian houses dated to the late Hellenistic period (mainly the late 2^nd to the early 1^st century BC), for example room AJ (service room) in the Maison des tritons and room c (probably a kitchen) in House II A in the Stadium Quarter18.

406Terra battuta/clay

4071. Trench A, stratum VII:

408A part of a burnt clay floor was located in the gap of the double wall, i.e. wall A1. The width is c.76 cm and the surface level -734 (Fig. 6).

4092. Trench B, stratum VI, eastern part of trench, east of wall B2:

410A hard clay layer containing fragments of tile and stone was registered at level -733/738 (Fig. 8).

4113. Trench B, stratum VII, western part of trench, between walls B1 and B2:

412The level for the layer is -750/752 (Fig. 8).

4134. Trench Ds, stratum X, west of wall D1, north of wall D2:

414The floor was partly dug through, but covered the area north of the transverse wall. The largest preserved surface was registered in the coordinates 81.11/124.04-81.0/125.0-82.65/124.29-82.65/124.89. In this area the surface level is -744/746 (Fig. 15).

4155. Trench K, stratum IV: The layer was registered in the western part of the trench and along the southern balk at level -657/668 (Fig. 52).

416Pavements not in situ

417Opus tessellatum

418Three white tesserae were found. Distribution: Ds IX (1); G II (1); G V (1).

419Mortar pavements19

420These pavements include a wide range of variants, but the predominant colours are: red/reddish (mixed with an aggregate of crushed tile or pottery); white (mixed with an aggregate of travertine); blackish-grey (mixed with an aggregate of lava). Mortar pavements may be undecorated or decorated. In the case of the latter, tesserae in stone or marble are set into the surface, scattered at random or arranged in geometric patterns. Moreover, larger fragments, again in stone or marble, may be set into the surface.

421The excavations yielded 11 fragments. The majority came from the area to the west of the street where they were found widely scattered both as regards trenches and strata (Table 5). One fragment was found in the zone occupied by the street (G IX), and only two came from the eastern trenches (L and N). The dimensions of the largest fragments are 6.5 cm x 5.3 cm and 6.7 cm x 4.3 cm. Generally, the mortar is white or whitish and mixed with terracotta chips of varying sizes.

422The scattered findspots and poor state of preservation, i.e. number and size, of the mortar fragments testify to a severe destruction of the pavements-whether their original contexts are within the excavated area (for which we have no evidence) or outside.

423The technique is documented as early as the late5^th and early 4^th century BC at Carthage20. In Sicily and on the Italian peninsula the earliest examples of mortar pavements, both with and without tessera decoration, date to the late 4^th and early 3^rd century BC, and they are extremely common in the late Republican period21. As far as Southern Italy is concerned, early pavements are documented at for instances Buccino (building at S. Stefano), Elea/Velia (bath) and Moltone di Tolve (habitation)22. Examples similar to the ones from Prop. Avallone occur at the Roman villa at Vittimose at Buccino where the floor in room 3 contains small brick particles (phase 2, i.e. Augustan period)23; and in House VII.7, 2 in Pompeii. Here a white mortar pavement is mixed with large terracotta fragments24. Rooms with cocciopesto floors are recorded both from the excavations in the Via Cavalleggeri situated south-east of Prop. Avallone and in Prop. Truono (cf. Fig. 2a). The former were abandoned in the late Republican period, while the latter seem to be of early Imperial date25. Unfortunately, these pavements are not described.

424CATALOGUE

425Cat. no. B-l: #89-204: Trench F. Stratum V. 85.15/136.67. Level -696. Fig. 65.

4266.5 x 5.0.

427Cat. no. B-2: #89-788: Trench G. Stratum IX. 88.5/126.0. Fig. 65.

4284.6 x 4.6.

429Floor tiles in terracotta26

430These are defined as thick flat tiles without flanges. The group consists of 16 specimens, all coming from upper layers dated to the Imperial period (Table 6). A common feature is a diagonal groove27.

431CATALOGUE

432Cat. no. B-3: #89-474: Trench E. Stratum IV. Fig. 68. Fragment with diagonal groove. Fabric: Reddish yellow (5 YR 6/6). 13.2 x 16.0; Th. 2.8.

433Cat. no. B-4: #90-1214: Trench H. Stratum II. Fig. 68. Fragment of comer with a faint groove along the edges. Fabric: Pinkish yellow (7.5 YR 8/2). 7.7 x 6.8; Th. 2.8.

434Wall decoration and pavements in stone: Opus sectile28

43516 fragments in marble and other stones may have been used for wall decoration and/or paving. The pieces were cut to specific shapes and fitted together to form a separate and smooth surface. Most fragments came from upper strata (Table 7).

II.3.2 Mudbricks

436The use of mudbrick for construction of superstructures/inner walls is only documented in con nection with the eastern insula and only in strata representing the first phase of occupation (Table 8). Its use, however, may also be conjectured for the contemporary structures west of the street, where postholes were found in trench A. The fragments were found in the immediate vicinity of the walls (cf. below). Moreover, mudbricks were registered, but not taken up, in K IV. A small amount came from trenches G and L. But most were discovered in trench H: of the 57 fragments found here all except one came from stratum V. An iron stud, cat. no. NN-35, was found in situ in one of the fragments from H V.

437In trench G the mudbricks were found in the area to the west of wall Gl, i.e. outside the building, while the ones from trench H came from both the area east and west of wall H1, i.e. inside and outside the building. Here they were secondarily burnt. In trench L the finds were concentrated in an area to the east of wall L1, 97.3-99.5/138.5-139.7, while one was found immediately to the west of the wall (#90-108a) and another on the wall (#90-108b).

438Generally, the clay is pink to reddish yellow (5 YR 7/4-7/8, 7.5 YR 7/4) on the outside with a dark grey (10 YR 4/1) core. The mudbricks are often poorly preserved, but a few larger fragments were recovered.

439CATALOGUE

440Cat. no. C-l: #89-801: Trench H. Stratum V. East of the wall. Destruction layer. 90.68/136.5. Fig. 66.

4414 fragments.

442Fabric: Reddish yellow (5 YR 7/8), greyish brown (10 YR 5/2) core. Largest fragment: 15.5 x 10.0 x 5.6.

443Cat. no. C-2: #89-809: Trench H. Stratum V. East of the wall. Destruction layer. 90.68/136.5. Fig. 67.

44417 fragments.

445Fabric: Reddish yellow (5 YR 7/8), grey (10 YR 5/1) core. Three largest fragments: 12.4 x 9.5 x 10.0; 10.0 x 9.0 x 8.0; 8.0 x 7.2 x 8.0.

II.3.3 The Roof Tiles

446Helle Damgaard Andersen and Birgit Tang

447The excavations yielded a huge amount of roof tiles represented by many types and fabrics29. The roof tiles include pantiles as well as cover and ridgepole tiles, of which it has been possible to classify a total of 1651 fragments (a fragment rates as one irrespective of size or the fact that it consists of several fragments put together). No complete tile is preserved.

448A problem when working with tiles is the lack of comparative material. This is due to the neglect of tiles in publications; only a few monographs on tiles in Antiquity exist. So far research has been done primarily in the field of Roman tiles, mainly in Britain30, and by Ö. Wikander in the field of Archaic Etruscan roof tiles31. However, in the case of Southern Italy there are publications that also include examples of ‘ordinary’ tiles with more or less detailed descriptions, e.g. from Monte Irsi and Pomarico Vecchio32.

449To a large extent the terminology established by Ö. Wikander is followed33, except that the sides of the pantiles are referred to as flanges34, not as raised borders. The pieces cut away at the end of the flanges, both on the top, outer side, and below the flanges are referred to as notches (by some scholars called cut-aways).

450The technique of cover tiles and ridgepole tiles is quite simple, since both are made from a flat piece of clay, moulded on a semi-circular wooden form35. As far as pantiles are concerned, they were made in moulds36. On several fragments of tiles (mostly pantiles) finger indentations or grooves can be seen, especially in the angle between the inside of the flange and the tile itself37.

451The following tables show the distribution of the different types of tiles. For ridgepole tiles, cover tiles and pantiles there is a general distribution table respectively, and additional tables present the distribution of each sub-type (cf. Tables 9-39).

452Almost all types of tiles were found in all trenches and all strata, thus relative chronology (not to mention absolute chronology) is uncertain. This signifies that roof tiles were reused to a large extent. It is not possible even to determine which type of cover tiles, pantiles, and ridgepole tiles were used together. However, some general patterns can be detected, indicating e.g. the locations of the ridges of the structures.

453For the ridgepole tiles, 22 fragments out of a total of 31 assignable fragments came from K V representing the first phase of occupation, i.e. the Courtyard House (Table 9). Types I and II occur in the first phase, and types II and III in the second phase.

454For cover tiles types I and III, H V and L IV provided the largest amounts, and together with K V these types clearly predominate in the eastern insula, meaning that they were used in the first phase of occupation (Table 13). However, all types are attested during this period. With type VI there is a clear change, as this type is most common in the western insula, thus being predominant in the second phase of occupation (B V, C V, Ds IX-X, E V, F V-VI). Taken as a whole, trenches D and E in the west and trenches H and L in the east yielded the largest amounts.

455As regards pantiles, it is characteristic that types 1.4 (high rounded profile) and IV (angular profile) with its sub-types, especially IV.1 and IV.2, are mainly found in the east (again the largest amounts came from the Courtyard House) (Table 23). Types 1.1-1.3 are most common in the west, i.e. in the second phase (B V, C V, Ds IX-X, E V).

456Due to the clear distribution of ridgepole tiles, it can be concluded that as far as the structure(s) in trenches D, E and F are concerned, the ridge must be situated west of the excavated area, or in the non-excavated areas between trenches D and E and between E and F. Another possibility is that cover tiles were employed for the ridge of the roof.

457CATALOGUE

458Ridgepole tiles

459Finds of ridgepole tiles are rare in Prop. Avallone. A total of 31 fragments were ascribed to three types based on a distinctive difference in fabric/clay (types I-III). By far the most documented is type I, whereas only a few fragments belong to types II and III (Table 9).

460Ridgepole tiles: type I

461The type represents semi-circular ridgepole tiles. On one fragment the entire length is preserved, 45.5 cm (#90-151 + #90-840 + #90-842-1 + #90-842-3 + #90-843-2 + #90-845 + #90-849). From this fragment it can be concluded that the type had one central opening for a cover tile on each side (Fig. 68). The coupling between the tiles was achieved by a flange (only one fragment has a flange preserved, #90-141) (Fig. 68). The height of the tiles is estimated at c.23.0 cm. The tiles are fairly thin, mostly less than 2.0 cm. The majority (at least 80%) came from trench K.

462Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 7/8), light grey (10 YR 7/2) core. Small and larger inclusions in black, brown and white, small stones and mica.

463Total ascribed: 27. Distribution: Table 10.

464Cat. no. D-l: #89-271 + #89-1027 + #89-635B + #89-635C: Trench H. Stratum V. East of the wall. Tile fall. 91.4/137.53. Level -680. Fig. 68.

465Fragment with part of one short side and opening for cover tile. L. 27.0; W. 24.0; est. H. c.23.0; Th. 2.0. The opening begins 16.5 cm from the short side and ends 23.5 cm from it.

466Cat. no. D-2: #90-141: Trench K. Stratum V. Tile fall. 95.12-95.30/137.6-137.7. Fig. 68.

467Fragment of short side with flange. L. 12.0; W. 16.5; Th. 1.8-2.2.

468Ridgepole tiles: type II

469The tiles belonging to this type are semi-circular in shape and fairly thick, between 2.6 and 3.5 cm.

470Fabric: Relatively coarse. Pink (5 YR 8/3). Yellow slip on top side. Many small and large inclusions in white, grey and red.

471Total ascribed: 3. Distribution: Table 11.

472Cat. no. D-3: #89-504: Trench Ds. Stratum X. West of the wall. Removal of tile fall. (Not illustrated) Fragment. L. 10.0; W. 13.4; Th. 2.6.

473Cat. no. D-4: #89-989: Trench G. Stratum IX. 88.5/126.0. (Not illustrated)

474Fragment with part of one short side. L. 6.6; W. 9.8; Th.>2.6.

475Ridgepole tiles: type III

476The type is characterized by its yellow fabric.

477Fabric: Coarse. Poorly levigated and secondarily burnt. Light red (2.5 YR 6/6). Small inclusions in grey and black, small stones and mica.

478Total ascribed: 1. Distribution: B V. Distribution: Table 12.

479Cat. no. D-5: #90-1757: Trench B. Stratum V. (Not illustrated)

480Fragment. L. 8.0; W. 11.5; Th. 2.3.

481Cover tiles

482Nine types of cover tiles were identified (types I-IX), comprising a total of 1113 fragments (Table 13). The classification of the types is based on shape, fabric, thickness and decoration. The largest groups of fragments belong to type I (313 fragments), type III (314) and type VI (308).

483Types I-III has an identical fabric, and they may all belong to the same type. Type I differs from type III in size by being lower. Type II is characterized by painted decoration.

484The majority of the cover tiles (71%) was found in trenches H (337 fragments, 285 from stratum V), L (238, 223 from stratum IV), E (110, 75 from stratum V), and Ds (105, 61 from stratum X).

485One example of a cover tile with a finger indentation has been preserved.

486Cover tiles: type I

487The type is characterized by simple thin semi-circular tiles (1.3 cm in thickness). The entire height and width of cat. no. D-6 is preserved: height 8.0 cm; width 14.5 cm. By far the largest amount of this type came from the Courtyard House.

488Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5YR 7/6). Many small inclusions in black, grey and white, a few larger ones, and mica. The fragments are often secondarily burnt. The fabric is identical to types II and III.

489Total ascribed: 313. Distribution: Table 14.

490Cat. no. D-6: #89-1150: Trench K. Stratum V. Tile fall. Fig. 68.

491Fragment with part of one short side and both long sides. L. 41.0; H. 8.0; W. 14.5; Th. 1.3.

492Cover tiles: type II

493The type is characterized by simple thin semi-circular tiles (1.1-1.6 cm in thickness). It differs from type I by its decoration with paint in black, brown, red and white/cream. c.80% of the fragments were found in H V.

494Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5YR 7/6). Many small inclusions in black, grey and white, a few larger ones and mica. The fragments are often secondarily burnt. The fabric is identical to types I and III.

495Total ascribed: 36. Distribution: Table 15.

496Cat. no. D-7: #89-803: Trench H. Stratum V. East of the wall. Removal of the tile fall. 90.68/136.5. Fig. 68.

497Fragment with red paint. L. 4.8; W. 7.3; Th. 1.4.

498Cat. no. D-8: #89-1041A-F: Trench H. Stratum V. East of the wall. 90.68/136.5. Fig. 68.

499A: Fragment with part of one long side and dark/black paint. L. 11.1; W. 5.0; Th. 1.5.

500B: Fragment with part of one short side and dark/black paint. L. 10; W. 5.7; Th. 1.4.

501C: Fragment with dark/black paint. L. 5.9; W. 3.7; Th. 1.4.

502D: Fragment with dark/black paint. L. 2.4; W. 2.8; Th. 1.1.

503E: Fragment with red paint. L. 3.6; W. 8.0; Th. 1.5.

504F: Fragment with white/cream paint. L. 6.0; W. 2.8; Th. 1.6.

505Cover tiles: type III

506The type is characterized by simple thin semi-circular tiles (1.1-1.6 cm in thickness). It is fairly tall. It differs from type I by being larger. On #89-1412 and cat. no. D-9 the entire height is preserved (9.3-10.5 cm as opposed to that of 8.0 cm of type I). Based on cat. no. D-9 the entire width is estimated at c.17.0 cm (type I 14.5 cm).

507Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5YR 7/6). Many small inclusions in black, grey and white, a few larger ones, and mica. The fragments are often secondarily burnt.

508The fabric is identical to types 1 and II.

509Total ascribed: 314. Distribution: Table 16.

510Cat. no. D-9: #90-1744: Trench K. Stratum V. Tile fall. Fig. 68.

511Fragment with part of one long side. L. 18.0; H. 10.5; Th. 1.3.

512Cover tiles: type IV

513The type is characterized by simple thin semi-circular tiles (1.3-1.6 cm in thickness). It is fairly tall. The entire height (or close to) has been preserved on cat. no. D-10, 9.6 cm.

514Fabric: Relatively well levigated and well fired. Reddish yellow (5 YR 7/6), pink (7.5 YR 8/4) on exterior. Many smaller and larger inclusions in white, brown, red and black, a little mica.

515Total ascribed: 74. Distribution: Table 17.

516Cat. no. D-10: #89-1664: Trench E. Stratum V. Fig. 68.

517Fragment with part of one long side. L. 16.0; W. 12.0; Th. 1.3.

518Cover tiles: type V

519The type is characterized by an angular profile. The tiles are thin (1.5-1.6 cm) and probably not very high (judging from cat. no. D-11). There are no traces of flanges.

520Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 7/6). Many small and larger inclusions in black and white, mica.

521Total ascribed: 7. Distribution: Table 18.

522Cat. no. D-ll: #90-780: Trench L. Stratum IV. Eastern part. 98.0/136.5. Fig. 68.

523Fragment with part of one short side. L. 16.5; W. 10.5; Th. 1.5.

524Cover tiles: type VI

525The type is characterized by simple semi-circular tiles. The thickness varies from 1.3-2.0 cm. Based on cat. no. D-12 the height is estimated at 8.5 cm, but the fragment #90-1772 may suggest that it was higher. The type is distinguished primarily by its fabric. The largest amount came from trenches D and E.

526Fabric: Relatively well levigated and well fired. Reddish yellow (5 YR 7/6), very pale brown (10 YR 8/4) slip. Many small-larger inclusions in brown, grey and red.

527Total ascribed: 308. Distribution: Table 19.

528Cat. no. D-12: #90-915: Trench F. Stratum VI. West of the wall. 82.5-83.5/136.5-139.5. Fig. 68.

529Fragment with part of one short and one long side. L. 28.0; H. 8.5; Th. 1.4.

530Cover tiles: type VII

531All the fragments are characterized by having flanges (or scars for flanges broken off). Based on the fabric, this type can belong to type I, II or III, or all of them, though the fairly small amount of type VII makes it unlikely that all three types had flanges. Based on #90-501, the tiles appear to be semi-circular and at least 15.0 cm in width-a width that may be common to all three types. The length of the flanges varies from 2.2 to 3.6 cm, but whether this variation is significant in relation to types is uncertain.

532Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 7/6). Many small inclusions in black, brown and white, and mica.

533Total ascribed: 24. Distribution: Table 20.

534Cat. no. D-13: #89-275: Trench H. Stratum V. East of the wall. Removal of the surface of the tile fall. Fig. 68.

535Fragment with flange. L. 10.5; W. 8.0; Th. 1.2.

536Cover tiles: type VIII

537The type consists of simple semi-circular tiles. The type is characterized by its fabric.

538Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 6/8). Numerous small inclusions in white, brown and red.

539Total ascribed: 7. Distribution: Table 21.

540Cat. no. D-14: #89-1655: Trench E. Stratum V. (Not illustrated) Fragment. L. 5.5; W. 10.0; Th. 1.2.

541Cover tiles: type IX

542The type consists of simple semi-circular tiles. The type is characterized by its fabric. Based on cat. no. D-15 its height can be estimated at 9.5 cm, and its width at around 13.5 cm.

543Fabric: Very well levigated and fairly well fired. Pink (5 YR 7/4). Small inclusions in black and white, mica.

544Total ascribed: 30. Distribution: Table 22.

545Cat. no. D-15: #90-1727: Trench A. Stratum III. Fig. 68.

546Fragment with part of one short and one long side. L. 22.5; H. 9.5; est. W. 13.5; Th. 1.6.

547Pantiles

548The excavations produced a huge amount of pantiles with and without flanges. Since it was not possible to determine the type of the tiles alone through fabric, only fragments with flanges have been included in the classification, comprising 507 fragments38.

549The pantiles have been divided into four main types according to the profile of the flange (types I-IV). Type I represents rounded profiles, high and relatively steep. Type II includes rounded but flatter profiles. Type III has oblique profiles and type IV angular ones. Each main type is subdivided because of a certain degree of variation, forming a total of 16 types and subtypes (Fig. 69).

550Type I is the largest group comprising 249 fragments, while type IV is the second largest (148). Type II includes 103 fragments and type III only 8.

551As shown by Table 23, the richest findspots for pantiles are as follows:

552Trench H (101, 74 from stratum V); trench Ds (87, 45 from stratum X); trench L (76, 74 from stratum IV), and trench E (66, 55 from stratum V).

553Pantiles: type I

554Type I represents rounded profiles and form the largest group. Rounded profiles are also present at Monte Irsi (Wightman 1977, 207, no. 372a, fig. 39).

555Total ascribed: 249.

556Pantiles: type 1.1

557The type is characterized by a high and relatively steep rounded flange. There is a notch on the top side and on the underside. There is no tapering. The entire width, 51.0 cm, is preserved on #89-176-8 + #89-176-16. The thickness varies from 1.9 to 3.6 cm. The height of the flange varies from 5.8 to 7.0 cm and the width of the flange from 4.3 to 6.0 cm. Approximately two-thirds of the tiles came from the trenches west of the street.

558Fabric: Well levigated and relatively well fired. Pink (5 YR 7/4), yellow (10 YR 8/4) slip. Sometimes the clay has a yellow exterior side. Many small and large inclusions in white, grey and brown. The fabric is similar to types II.3 and II.4, fabric B.

559Total ascribed: 83. Distribution: Table 24.

560Cat. no. D-16: #89-1415: Trench H. Stratum V. East of the wall. Tile fall. 90.68/136.5.Level -672. Fig. 69.

561Fragment with notch on the lower side. L. 10.5; W. 8.0; Th. 2.8; H. of flange 6.4; W. of flange 5.3. L. of notch 7.0.

562Pantiles: type 1.2

563The type is similar to type 1.1, though often with a slightly lower flange. There is a notch on the top side and underside. As opposed to type 1.1, the tile is tapering. The thickness varies from 2.0-3.1 cm. The height of the flange varies from 5.7-6.6 cm and the width of the flange from 3.7-5.3 cm.

564Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 7/6), traces of yellow slip. Many small and large inclusions in white, red, brown and black.

565Total ascribed: 51. Distribution: Table 25.

566Cat. no. D-17: #89-1661: Trench Ds. Stratum X. Fig. 69.

567Fragment of upper left corner with notch on the top side. L. 16.5; W. 8.5; Th. 2.1; H. of flange 6.0; W. of flange 3.8; L. of notch 2.5.

568Pantiles: type 1.3

569The type is characterized by a lower flange, which is often fairly flat on the top. There is a notch on the top side and underside. There is no tapering. The thickness varies from 2.5-3.3 cm. The height of the flange varies from 5.2-5.9 cm and the width of the flange from 3.8-5.3 cm.

570Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 7/6), very pale brown (10 YR 8/3) slip (#174-9-297). Many small and lager inclusions in white, grey, brown and red, mica. Dark paint can be seen on #90-1730.

571Total ascribed: 22. Distribution: Table 26.

572Cat. no. D-18: #89-176-6: Trench E. Stratum V. Figs. 69 and 71.

573Fragment of lower right comer with notch on the underside. Print of a child’s hand on the underside. L. 31.5; W. 39.0; Th. 3.3; H. of flange 5.5; W. of flange 4.6; L. of notch 7.5.

574Pantiles: type 1.4

575The type is characterized by rounded, but slightly oblique flanges. There is a notch on the top side and on the underside. There is no tapering. The thickness varies from 1.9-3.3 cm. The height of the flange varies from 4.5-6.3 cm and the width of the flange from 3.7-5.2 cm.

576Fabric: The type has two different fabrics, fabric A being the most common. Both fabrics have notches on the top side and underside.

577Fabric A: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 7/8). Small and larger inclusions in black, red, brown and white, mica. Almost all the fragments have been secondarily burnt.

578Fabric B: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 7/6), very pale brown (10 YR 8/3) slip. Small and large inclusions in grey, red and brown, mica. Only two fragments represent this fabric, including cat. no. D-19.

579Total ascribed: 67. Distribution: Table 27.

580Cat. no. D-19: #89-1670 (fabric B): Trench E. Stratum V. West of the wall. 83.5-84.5/129.5. Fig. 69.

581Fragment of upper right corner with notch on the top side. L. 11.0; W. 9.5; Th. 2.2; H. of flange 5.5; W. of flange 5.2; L. of notch 2.0.

582Pantiles: type 1.5

583The type is characterized by a fairly short rounded profile. There is a notch on the top side and on the underside. There is no tapering. The thickness varies from 2.0 to 2.7 cm. The height of the flange varies from 4.8 to 6.4 cm and the width of the flange from 3.1 to 4.6 cm.

584Fabric: The type has two different fabrics. Notches are only documented on the top side on fragments of fabric A, while for fabric B notches are documented both on the top side and the underside.

585Fabric A: Well levigated and relatively well fired. Very pale brown (10 YR 8/4). Small and larger inclusions in grey, brown and red.

586Fabric B: Relatively well levigated and relatively well fired. Secondarily burnt. Reddish yellow (5 YR 7/8), yellow/cream slip (#89-1715, #90-1758).

587Total ascribed: 18. Distribution: Table 28.

588Cat. no. D-20: #90-1732 (fabric A): Trench E. Stratum VI. East of the pavement. 82.5/129.68. Fig. 69.

589Fragment of upper right comer with notch on the top side. L. 10.0; W. 9.5; Th. 2.0; H. of flange 5.5; W. of flange 3.3; L. of notch 1.6.

590Pantiles: type 1.6

591The type is characterized by a short rounded, but slightly oblique profile. Notches are only documented on the top side. There is no tapering. The thickness varies from 2.4 to 2.8 cm. The height of the flange varies from 4.9 to 6.1 cm and the width of the flange from 3.2 to 3.8 cm.

592Fabric: Relatively well levigated and relatively well fired. Pink (5 YR 7/4), yellow on the top side, very pale brown (10 YR 8/3) slip. Small and large inclusions in brown, grey and red.

593Total ascribed: 8. Distribution: Table 29.

594Cat. no. D-21: #174-9-306: Trench Ds. Stratum IX. Fig. 69.

595Fragment of upper left corner with notch on the top side. L. 18.5; W. 14.5; Th. 2.5; H. of flange 4.9; W. of flange 3.5; L. of notch 1.8.

596Pantiles: type II

597Type II represents rounded profiles like type I, but they are flatter.

598Total ascribed: 103.

599Pantiles: type II. 1

600The type is characterized by a large rounded, fairly flat profile. Notches are only documented on the underside. There is no tapering. The thickness varies from 2.9-3.4 cm. The height of the flange varies from 5.6-6.4 cm and the width of the flange from 6.1-7.5 cm. It is comparable to Monte Irsi (Wightman 1977, 207, no. 381, fig. 40, from a layer predating the Roman building on the site).

601Fabric: The type is represented by two different fabrics

602Fabric A: Relatively well levigated and relatively well fired. Reddish yellow (7.5 YR 7/6). Many inclusions in red, black and white, mica.

603Fabric B: Relatively well levigated and relatively well fired. Very pale brown (10 YR 7/4). Large inclusions, stones, and air pockets. This fabric is attested by only one fragment (#89-1710).

604Total ascribed: 7. Distribution: Table 30.

605Cat. no. D-22: #90-1761 (fabric A): Trench C. Stratum V. Fig. 70.

606Left side with notch on the underside. L. 14.5; W. 12.0; Th. 3.2; H. of flange 6.0; W. of flange 7.0.

607Pantiles: type II.2

608The type is characterized by a fairly small rounded, fairly flat profile. It is provided with notches on the underside. There is no tapering. The width is 46.0-46.5 cm (entirely preserved on cat. nos. D-24 and D-25). The thickness varies from 2.1 to 3.0 cm. The height of the flange varies from 4.1 to 5.5 cm and the width of the flange from 3.7 to 5.5 cm.

609Fabric: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 6/6). Many small inclusions in white and red, stone and mica. The fabric is identical to type II.4, fabric A.

610Total ascribed: 78. Distribution: Table 31.

611Cat. no. D-23: #89-1402: Trench G. Stratum VIII. Fig. 70.

612L. 22.0; W. 13.0; Th. 2.7; H. of flange 5.0; W. of flange 5.0.

613Cat. no. D-24: #90-917: Trench J. Stratum V. Tile fall. Lower stratum, c.90.7/145.5. Fig. 71.

614Lower part with notch in both sides on the underside. The width is entirely preserved. L. 32.5; W. 46.0; Th. 2.5; H. of flange 4.1; W. of flange 5.2; L. of notch in left side 10.5; L. of notch in right side 10.5.

615Cat. no. D-25: #90-1200: Trench J. Stratum V. Tile fall. Lower stratum, c.90.7/145.5. Level -709. Fig. 71.

616Lower part with notch in both sides on the underside. The width is entirely preserved. L. 36.5; W. 46.5; Th. 2.6; H. of flange 5.0-5.5; W. of flange 4.5; L. of notch in left side 11.5; L. of notch in right side 8.0.

617Pantiles: type II.3

618The type is characterized by low, flat, slightly rounded profiles. Only the underside has notches. There is no tapering. The thickness varies from 3.1 to 3.3 cm. The height of the flange varies from 4.9 to 5.4 cm and the width of the flange from 4.1 to 5.2 cm. The length of this type exceeds 49.5 cm and the width 32.5.

619Fabric: Well levigated and relatively well fired. Pink (5 YR 7/4). Many small and large inclusions in white, grey, and brown. The fabric is similar to types 1.1 and II.4, fabric B.

620Total ascribed: 7. Distribution: Table 32.

621Cat. no. D-26: #90-1215: Trench J. Stratum IV. 89.5-90.2/145.4-147.0. Fig. 70.

622Upper right comer. One finger indentation. L. 26.5; W. 13.0; Th. 3.1; H. of flange 4.9; W. of flange 4.1.

623Pantiles: type II.4

624The type is characterized by a rounded profile similar to type II.2, but slightly more oblique. A notch is documented on the top side, and there was probably also a notch on the underside. There is no tapering. The thickness varies from 1.8-2.7 cm. The height of the flange varies from 4.2-5.0 cm and the width of the flange from 3.6-4.4 cm.

625Fabric: The type has two fabrics.

626Fabric A: Relatively well levigated and relatively well fired. Reddish yellow (5 YR 6/6). Many small inclusions in white and red, stone and mica. The fabric is identical to type II.2

627Fabric B: Well levigated and relatively well fired. Pink (5 YR 7/4). Many small and large inclusions in white, grey, and brown. The fabric is identical to types 1.1 and II.3.

628Total ascribed: 11. Distribution: Table 33.

629Cat. no. D-27: #89-1699: Trench E. Stratum V. Level -688. Fig. 70.

630L. 13.0; W. 14.0; Th. 2.5; H. of flange 5.0; W. of flange 3.9.

631Pantiles: type III

632Type III has oblique profiles. These occur at Monte Irsi (Wightman 1977, 207, no. 372, fig. 39).

633Total ascribed: 8.

634Pantiles: type III. 1

635The type is characterized a fairly large and oblique profile. Notches are found on the underside. The thickness varies from 1.4-2.6 cm. The height of the flange varies from 3.9-6.2 cm and the width of the flange from 3.4-4.9 cm.

636Fabric: Relatively poorly levigated and relatively badly fired. Coarse. Reddish yellow (5 YR 6/6), grey core. Many inclusions in black, white and red, small stones.

637Total ascribed: 7. Distribution: Table 34.

638Cat. no. D-28: #90-1721: Trench H. Stratum V. Tile fall. 89.5/140.0-142.5. Fig. 70.

639L. 17.0; W. 14.5; Th. 2.6; H. of flange 6.2; W. of flange 4.5.

640Pantiles: type III.2

641This type is represented by one fragment only, but since it is so distinctive, it has been given its own type number. The type is characterized by being extremely thin with a small oblique profile.

642Fabric: Well levigated and well fired. Pink (5 YR 7/4), very pale brown (10 YR 8/3) surface and slip. Many small black and white inclusions.

643Total ascribed: 1. Distribution: F IV. Distribution: Table 35.

644Cat. no. D-29: #89-559: Trench F. Stratum IV. Fig. 70.

645L. 11.2; W. 10.2; Th. 1.6; H. of flange 2.7; W. of flange 2.0.

646Pantiles: type IV

647Type IV represents tiles with an angular profile. Again parallels are found at Monte Irsi (Wightman 1977, 208, no. 376, fig. 39).

648Total ascribed: 148.

649Pantiles: type IV. 1

650The type is characterized by an angular profile. There is a notch on the top side, but instead of a notch on the underside there is an indentation in the flange. Both the tile and the flanges markedly taper. Based on the preserved fragments, the length can be estimated as approximately 60.0 cm. The thickness varies from 2.6 to 3.3 cm. The height of the flange varies from 4.9 to 6.1 cm and the width of the flange from 1.8 to 4.5 cm.

651Fabric: Fairly well levigated and medium fired. Reddish yellow (7.5 YR 7/6), grey core. Many small inclusions in black and white, mica. Several fragments have a cream slip on the top side and some also on the underside. Several fragments are decorated with dark paint (black? red, and brown). Some fragments have been secondarily burnt.

652Total ascribed: 56. Distribution: Table 36.

653Cat. nb. D-30: #89-1718: Trench H. Stratum V. East of the wall. Removal of the tile fall. Level -672.90.68/136.5. Fig. 70.

654Fragment of upper right comer. L. 14.3; W. 8.6; Th. 2.7; H. of flange 5.0; W. of flange 1.8-2.1.

655Pantiles: type IV.2

656The type is characterized by a simple angular profile. A notch is only documented on the underside. There is no tapering. The thickness varies from 1.6 to 2.8 cm. The height of the flange varies from 4.8 to 6.2 cm and the width of the flange from 2.5 to 3.5 cm.

657Fabric: Very coarse and badly fired. Reddish yellow (5 YR 6/6), grey core. Many small to large inclusions in white and brown, mica. Dark paint (originally red) on two fragments.

658Total ascribed: 61. Distribution: Table 37.

659Cat. no. D-31: #90-1746: Trench K. Stratum V. Tile fall. Fig. 70.

660L. 29.4; W. 27.0; Th. 2.6; H. of flange 4.8; W. of flange 3.0.

661Pantiles: type IV.3

662The type is characterized by a simple angular profile, slightly rounded at the top. Notches are only documented on the underside. The thickness varies from 1.8 to 3.0 cm. The height of the flange varies from 4.8 to 5.9 cm and the width of the flange from 2.6 to 4.3 cm. One fragment was found in Tomb 6093 where it functioned as a cover. The profile of the flange is more oblique than the other sub-types.

663Fabric A: #90-1728: Relatively well levigated and well fired. Reddish yellow (5 YR 7/6). Small inclusions in black, white, a little mica.

664Fabric B: #90-1217: Relatively well levigated and well fired. Pink (5 YR 7/3). Small and larger inclusions in white and brown.

665Fabric C: #89-1712: Coarser clay. Relatively well fired. Reddish yellow (5 YR 6/6). Many small and larger inclusions in white and dark colours.

666Total ascribed: 20. Distribution: Table 38.

667Cat. no. D-32: #89-1712 (Fabric C): Trench H. Stratum V. West of the wall. Level -675. Fig. 70.

668L. 14.5; W. 8.2; Th. 2.1; H. of flange 5.1; W. of flange 2.6.

669Cat. no. D-33: Without no.: Trench J. Stratum III. Tomb 6093. Fig. 70.

670White slip on top and inside of the flange. L. 34.0; W. 26.8; Th. 2.3; H. of flange 5.3; W. of flange 3.7.

671Pantiles: type IV.4

672The type is characterized by a simple angular profile, slightly rounded at the top; similar to type IV. 3, but thicker. Notches are only documented on the underside. The thickness varies from 2.3 to 2.7 cm. The height of the flange varies from 5.5 to 6.2 cm and the width of the flange from 3.4 to 4.4 cm.

673Fabric: Relatively well levigated and well fired. Light reddish brown (5 YR 6/4), very pale brown (10 YR 8/4) slip. Small and larger inclusions in light and dark colours, a little mica.

674Total ascribed: 10. Distribution: Table 39.

675Cat. no. D-34: #89-1708: Trench E. Stratum V. Level -688. Fig. 70.

676Yellow slip on top side. L. 19.1; W. 7.6; Th. 2.4; H. of flange 6.2; W. of flange 4.4.

II.3.4 Miscellaneous tiles

677Helle Damgaard Andersen and Birgit Tang

678Several fragments of tiles are decorated, have marks or prints. The best preserved are listed in the catalogue.

679CATALOGUE

680Stamped decoration

681Cat. no. D-35: #89-543: Trench Ds. Stratum X. West of the wall. Removal of the tile fall. Fig. 70.

682Fragment with rosette.

683Fabric: Reddish yellow (5 YR 6/6). 22.5 x 14.5; Th. 3.0.

684Cat. no. D-36: #89-21: Trench E. Stratum IV. 82.49/129.97. Level -630. Fig. 70.

685Fragment with illegible, rectangular stamp.

686Fabric: Reddish yellow (7.5 YR 7/6). 11.3 x 10.5; Th. 3.7.

687Cat. no. D-37: #89-915: Trench H. Stratum V (probably fallen from stratum III). West of the wall. 88.5-91.5/136.5138.5. Fig. 70.

688Fragment with letters: [...MEL...].

689Fabric: Pink (5 YR 7/4). L. 6.9; W. 4.5; Th. 3.0.

690Possible marks or letters

691Cat. no. D-38: #89-90: Trench E. Stratum III. 84.5/ 129.5. Level -531. Fig. 70.

692Fragment of pantile with crossed diagonal grooves (letter?). 5.7 x 5.6.

693Cat. no. D-39: #89-1666: Trench E. Stratum V. Level -688. Fig. 70.

694Fragment of pantile with a groove in an angle (letter?). 15.4 x 9.4; Th. 3.3

695Cat. no. D-40: #89-1669: Trench E. Stratum V. West of the wall. 83.3-84.5/129.5. Fig. 70.

696Fragment of cover tile with a straight ridge (a mark/letter?). 18.4 x 10.1.

697Painted decoration

698Some fragments of pantiles without flanges preserve traces of painted decoration, consisting of dark paint and cream paint combined with dark paint (#89-807).

699Cat. no. D-41: #89-807: Trench G. Stratum VIII. 85.0-87.5/126.0-128.0. Fig. 71.

700Fragment with dark/black and cream/white paint.

701Fabric: Reddish yellow (5 YR 7/6). 6.3 x 5.1; Th. 2.7.

702Cat. no. D-42: #89-916: Trench H. Stratum V. East of the wall. 88.5-91.5/136.5-138.5. Fig. 71.

703Fragment with dark paint (assignable to type IV. 1 due to the fabric). 5.7 x 5.5; Th. 3.1.

704In addition to cover tiles type II, there are various unassignable fragments of cover tiles with remains of paint or slip.

705Animal and human prints

706Nine tiles with footprints of animals were found and three with human prints, of which two have fingerprints, and one a print of a child’s hand (cat. no. D-18).

707Cat. no. D-43: #89-425: Trench Ds. Stratum X. East of the wall. Removal of the tile fall. Fig. 70.

708Fragment of a pantile with an imprint of an animal’s paw.

709Fabric: Reddish yellow (5 YR 6/6). 18.5 x 13.5 x 2.2.

710Cat. no. D-18: #89-176-6: Trench E. Stratum V. Fig. 71.

711Fragment of pantile with flange, lower right comer with notch on the underside. Print of a child’s hand on the underside, (type 1.3).

712Fabric: Reddish yellow (5 YR 7/6). L. 31.5; W. 39.0; Th. 3.3; H. of flange 5.5; W. of flange 4.6; L. of notch 7.5.

713Cat. no. D-44: #89-458: Trench E. Stratum V. Below the tile fall. 84.5/129.5. Fig. 70.

714Fragment of pantile with footprint of animal.

715Fabric; Reddish yellow (5 YR 7/6).

71612.0 x 10.0; Th. 2.0.

717Cat. no. D-45: #90-692: Trench L. Stratum III. 98.5-99.5/136.5-137.2. Fig. 71.

718Fragment of pantile with flange and notch on underside. Footprint of an animal on top side.

719L. 14.8; W. 22.5; Th. 2.7; H. of flange 5.5; W. of flange 4.0.

II.3.5 Architectural terracottas

720Helle Damgaard Andersen and Birgit Tang

721Three fragments of revetments were found, among which cat. no. E-2 can be identified as a raking sima. Cat. no. E-l is decorated with a moulded floral design, and E-2 carries a simple linear design.

722CATALOGUE

723Cat. no. E-l: #89-8: Trench Ds. Stratum X. Fig. 70.

724Fragment of a revetment plaque with a guilloche design in relief, bordered at least on one side by a horizontal groove.

725Fabric: Pink (7.5 YR 8/4). 11.5 x 10.7; Th. 3.0

726Cat. no. E-2: #89-22: Trench E. Stratum IV. Fig. 70.

727Fragment of the lower part of a raking sima. On the protmding lower part is a moulded design consisting of a series of three thin vertical grooves, separated by an undecorated space.

728Fabric: Reddish yellow (7.5 YR 7/6). H. 4.6; 18.5 x 10.5; Th. 3.3.

II.3.6 Water pipes

729Helle Damgaard Andersen and Birgit Tang

73012 fragments of water pipes were found in layers of the Imperial period; none of them were in situ (Table 40). The clay is identical (reddish yellow, 5 YR 6/6), and the estimated exterior diameter is 18.0 cm. Flanges interconnected the pipes (Fig. 70). Except for one fragment (#90-1592), they all have a layer of lime on the inside. Furthermore, a lime-sediment with the print of a water pipe came from H IV.

731CATALOGUE

732Cat. no. F-l: #174-5-53: Trench Ds. Stratum V. Fig. 70.

733Fragment of finished edge with flange. L. 6.8; W. 9.0; Th. 0.8; L. of flange 4.4; Th. of flange 1.0.

734Cat. no. F-2: #90-1590: Trench J. Stratum III. Tile fall. Fig. 70.

735Fragment of finished edge with flange. L. 7.0; W. 7.0; L. of flange 5.6; Th. of flange 1.0

II.4 The tombs*

736Birgit Tang

737Three child burials in amphorae (ad enchytrismoi) were found in stratum III in the northernmost part of the excavated area (not including trench N). The tombs are identified by their numbers in the Pontecagnano sequence: Tombs 6008, 6092 and 6093. Tomb 6008 appeared while a bulldozer was peeling off the upper layers prior to the opening of new trenches in the extension of the excavation area in 1989. Tombs 6092 and 6093 were both excavated in 1990. Apart from these three well-defined burials, a cluster of human bones were found in H IV (designated ‘Suspected Tomb’ in the analysis of the skeletal remains). Moreover, remains found in F IV are likely to derive from a fourth child burial. These remains consist of large fragments of an amphora and bones39. The location of the three tombs is shown in Fig. 72. The section on the amphorae is followed by the analysis of the skeletal remains.

738Due to their location within a confined area and at the same level (-629.5, -627/637, -625), the three enchytrismoi presumably represent a limited period. The most probable date for all three types of amphorae is the 3^rd century AD, and more likely the first half or the mid-3^rd century than the second half, as Gauloise 4 amphorae (cat. no. G-2) are rare after the mid-3^rd century AD. Although the Iberian amphora (cat. no. G-l) from Tomb 6008 is mainly a product of the 4^th to5^th centuries AD, it is attested in Ostia in the first half of the 3^rd century AD. Within the same century the vessels were reused as burial containers for newborn children. Water had run through the amphorae in Tombs 6008 and 6093 carrying diminutive glass fragments with it (cf. Marshall Becker’s analysis of the skeletal remains). Possibly the area was used also as burial ground for older children, since some bones of a 14/17-year-old person were recovered in H IV, i.e. just above the building (the so-called Suspected Tomb).

739As far as chronology is concerned, it can be concluded that the amphora burials disturbing stratum III give a terminus ante quem of the 3^rd century AD, at the latest the mid- 3^rd century, for the layer, at least in this particular zone. The amphora types also testify to the variation of imported products in the area.

II.4.1 The amphorae40

740Tomb 6008 (Fig. 73)

741Cat. no. G-l: Trench H. Stratum III. 91.17-91.45/137.13-137.84. Level -629.5.

742The amphora belongs to the type Peacock & Williams Class 23/Almagro 51c/Keay XXIII/Augst 40Gruppe 13) characterized by a triangular-shaped rim to which two broad handles are attached and provided with a groove down the centre, a narrow, conical neck, a high-shouldered piriform body and a short hollow toe. The variations and sub-types, however, are numerous, and a linear typological development cannot be established41. The dimensions of the Prop. Avallone amphora are close to the most typical examples from the Catalan material analyzed by Keay; they are 74.1 cm high and have a maximum diameter of 33.7 cm42. The fabric seems comparable to fabric TG 65 at Augst43. This type of amphora was produced in several workshops on the Iberian peninsula, e.g. in Baetica in the 3^rd to the beginning of the 5^th century AD, and served as container for various products, primarily fish sauce and salted fishes44. In Ostia the earliest contexts belong to the first half of the 3^rd century AD, but the main period is the 4^th to 5^th centuries AD45.

743References; Manacorda 1977, 142-145 and 278, cf. 255, fig. 4; Keay 1984, 172-178; Peacock & Williams 1991, 132-133; Martin-Kilcher 1994, vol. 2, 430-431.

744The amphora was found in several fragments, damaged by the bulldozer; the lower part of the burial was, however, still in situ, comprising the lower part of the neck with one handle, about half the body and the toe with the skeletal remains inside. Presumably, a large part of the body of the vessel was cut away in order to place the body of the child, and possibly a fragment of a roof tile served as lid (as in the case of Tomb 6093, cf. below), so that this part of the amphora was missing already in antiquity.

745Fabric: Very coarse. Very pale brown (10 YR 8/3) outer surface. Light red (close to 10 R 6/8, but still lighter) core, some white mica and a few larger, brown particles.

746H. 68.0; max. 0 31.0; Th. of wall 1.0.

747Tomb 6093 (Fig. 73)

748Cat. no. G-2: Trench J. Stratum III. c. 89.5-90.12/145.92-146.32. Level -625 (bottom level).

749The amphora is a Gauloise 4/Pélichet 47/Peacock & Williams Class 27 defined by its thick and rounded rim, short neck, broad shoulders, thin-walled piriform body, and a narrow flat base.

750This wine amphora was primarily produced in southern France, but local imitations were also manufactured in Baetica in the 3^rd century AD46. It is dated to the mid- 1^st to 3^rd century AD. In Ostia the finds are rare beyond the mid-3^rd century AD47.

751References: Panella 1973, 538-551; Manacorda 1977, 122-145 and 149; Peacock & Williams 1991, 142-143; Laubenheimer 1985, 261-293 and 400-403; Laubenheimer 1990, especially 98-104; Laubenheimer 1998, 169-171.

752Here too the amphora was lying in an almost east-west position with its mouth towards the west. A large piece of the body of the amphora had been removed for the interment of the child, and the hole was covered with a pantile (type IV.3). Consequently, the upper part with rim, neck and handles are missing. Only the remnants of one of the presumably flat? handles are preserved, but broken so that its width is not determinable. The wall of the amphora is somewhat depressed. Horizontal grooves are present on the upper part of the body. The earth fill of the amphora did not contain any ceramic fragments, but three blackgloss sherds were found below the amphora.

753Fabric: Pinkish white to pink (7.5 YR 8/2-8/4), much, especially brown and black, mica.

754H. 62.0; max. 0 38.0; Th. of wall 1.1.

755Tomb 6092 (Fig. 73)

756Cat. no. G-3: Trench L. Stratum III. 98.95-99.97/139.75-140.17. Level -627/637 (spike/mouth).

757The amphora belongs to the type Africana I/Africana Piccola/Peacock & Williams Class 33/Keay III/Augst 65 provided with a thickened ring-like rim, a short, slightly conical neck and a long cylindrical body with a short spike. Two short rounded handles with a central groove are attached to neck and shoulder. The neck has a rectangular depression with the stamp:

758TOP

759DVMVR

760The TOP series occurs on both Africana I (Africana Piccola) and Africana II (Africana Grande) amphorae, e.g. in Ostia (cf. below). Comparanda, though not identical, for the stamp are found in Rome and Ostia:

7611) Rome: Africana I amphora from the Esquiline, CIL XV, 552-553, No. 3537, post-Severian48:

762TOPOL/[...]VMVIRI

7632) Ostia: Africana II amphora: Manacorda 1977, 195, No. 104:

764TOP/DVMR (?) or more likely MVR judging from Tomb 6092 and the Esquiline inscription.

765The type was produced in North Africa, Tunisia and used for olive oil49. It is documented in contexts from the second half of the 2^nd century AD (e.g. Ostia and Tipasa) to the 4^th century AD. The majority of the fragments from both Monte Testaccio in Rome and Ostia belong to the 3^rd century BC50.

766References: Panella 1973, 574-579; Manacorda 1977, 124, 156-159; Peacock & Williams 1991, 153-154; Martin-Kilcher 1994, vol. 2, 444-448; Bonifay 2004, 106-107.

767The amphora was found in situ, but was crushed into several fragments with its mouth towards the west. One handle is missing. In order to place the body of the child inside it, the amphora was horizontally broken into two parts just below the shoulder by a series of short chisel strokes, so that a few centimetres are missing on one side. Apart from the skeletal material, the fill of the amphora contained a few ceramic fragments, none of which contribute to its chronology.

768Fabric: Light red to red (10 R 6/6-5/6, but slightly darker), much small, especially white, mica. Pinkish grey (7.5 YR 6/2) slip/wash?

769H. 104.0; 0 29.0; Th. of wall 1.2; Th. of rim 2.4.

II.4.2 The anthropological analysis of the human skeletal remains: Tombs 6008, 6093 and 6092. Suspected Tomb

770Marshall Joseph Becker, Ph.D.51

771Professor of Anthropology, West Chester University of Pennsylvania

772The actual amphorae containing the three enchytrismos burials were filled with earth which had seeped into each over the centuries. Thus each of these vessels had to be ‘excavated’ in the laboratory. In addition to the narrow necks of these vessels a hole had been made in the body of each to enable the body of the infant, probably wrapped in some kind of protective covering, to be introduced into the container. Skeletal preservation of these three infants was directly proportioned to the condition of the vessels within which they were found. Remarkable preservation was noted for the bones from within the amphora which was almost intact, but even the two damaged vessels contain bone in sufficiently good condition to permit an extremely exact evaluation of age to be made. Although the determination of gender for foetal remains is not considered to be reliable, figures calculating foetal length have been used to make estimates52. For lengths and midshaft diameters of long bone shafts, see Table 41.

773Tomb 6008 (Figs. 48 and 73)

774Trench H. Stratum III. 91.17-91.45/137.13-137.84. Level -629.5

775Age: in utero 8 months (stillborn or died shortly after birth).

776The amphora containing the remains of this child may have been damaged by bulldozing operations over this area of the archaeological excavation. However, the distribution of the tiny bones suggests that some ancient disturbance may have had greater effects on the condition of these remains. The amphora was found lying on its side and the bones were concentrated in the lowest point, which was at the widest point of the vessel (at the shoulder, just below the neck). Although the positions of these bones were generally indicative of an articulated skeleton, several had been clearly relocated in antiquity. The ancient disturbance may have taken place while the chamber formed by the vessel was still unclogged with the earth which subsequently seeped in to fill the entire cavity. No layering of the sandy fill is evident. Stones and sherds up to 15 mm in diameter had also filtered into the vessels along with the earth.

777Quite clearly the child had been placed within this vessel with its head toward the original opening and its feet toward the narrow base. However, the unfused left frontal bone was found near the legs. The right femur was found fused to the innominates by a calcareus deposit, probably resulting from flooding of this chamber at some point or points after the small corpse had completely decomposed. Since this process would have taken but a few weeks or possibly a few months depending on the season of the year, the wet environment leading to the formation of this deposit may have been generated at almost any time within the first century after burial. Once the vessel began to fill with earth, such deposits would have formed elsewhere, such as on the surface of the intruded fill.

778Also important in evaluating this calcareous deposit is its presence only around the bones in the middle of the skeleton. This reflects the position of the body in this tiny chamber, with its hips at the lowest point, which is where water would collect and where lime deposits would be most likely to be found.

779The three shafts of the major bones of the left arm are present and indicate that they were flexed. However, all three bones of the right arm as well as the left tibia are missing.

780Dentition is represented by only deciduous teeth. What appear to be both left maxillary incisors were found in place within a section of decomposing bone, but only the very tips have formed. Both are pointing away from the orbit, indicating that they had come into proper alignment for eruption. What appears to be d1I is 6.5 mm board and nearly 4 mm high, while the smaller is c. 5 mm broad and 1 mm high.

781Both incisors of the right mandible also are in place. The dI1 is lying in a horizontal position with only 3/5 of the crown complete, and the dI2 has a crown which is half complete. The right deciduous canine tip is in place. The first molar may be represented by a tiny fragment of crown found loose. Three cusps of a molar crown in this mandibular fragment (dM2?) were found fused, and a fourth cusp found loose may belong with this tooth. Since these are the only facial parts and teeth recovered, the missing portions (like the missing long bone shaft) reflect the degree to which later disturbances have influenced the recovery of skeletal material from this context, when compared with recovery from the almost intact amphora of Tomb 6093 (see below).

782Note also should be made of a small number of glass fragments, all of which appear to be under 5 mm at their greatest length. Each of these tombs has glass in association, with the size of the fragments perhaps indicative of transport patterns or distance from source.

783Tomb 6093 (Figs. 51 and 73)

784Trench J. Stratum III. 89.5-90.12/145.92-146.32. Level -625 (bottom level).

785Age: 8 months in utero.

786As with the other vessels, the entry hole for inserting the infant was made at the widest part, at the high shoulder of this amphora. The head of the child was placed toward the neck/opening and the feet toward the toe of the vessel. The vessel is almost intact, and the skeleton within has similarly survived in excellent condition and nearly perfect articulation. The child is lying in an extended supine position, with hands in the lap area. The left frontal, however, appears by the legs, possibly having floated there during a wet period after earth had buried (and anchored) most of the very small bones beneath a thin layer of soil. A complete inventory of these tiny elements is not needed, but almost every bone of this infant has been recovered, including the right and left malleus (of the ears).

787The left mandible was found intact, but the five deciduous teeth and whatever aspects of the 1M which had developed all had fallen out. Only the dlM was found loose, with 3/4 of the crown surface formed and fused. The right mandible, on the other hand, is intact and has within it three of the first four teeth. The dI2 is represented only by the tip of the crown deep in its socket and pointing down away from the alveolar edge. While the dC is not in place, or too little developed to be detected, the dMl is represented by a fused cusp surface, still inverted in its socket, representing the entire tip of the crown with the exception of a few open areas at the anterior aspect. The dM2 and Ml were not seen.

788Although all sockets of both maxillae are represented, the bones are not intact and none of the teeth were found in situ. However, both maxillary central incisors have been found loose, with the crowns 1/2 to 3/5 complete. Both deciduous first molars were also recovered, with the crown surface complete on the right, but only nearly complete for the left with the exception of a small hole. Thus development/eruption for the left maxilla appears somewhat slower than the right.

789Fewer glass fragments were found in the fill within this amphora than in the fill from Tomb 6008, but all seem of slightly larger size (c.3-6mm).

790Tomb 6092 (Figs. 56 and 73)

791Trench L. Stratum III. 98.95-99.97/139.75-140.17. Level -627/637.

792Age: 4.5 + 1.5 months.

793The bones within this amphora had been seriously disturbed in antiquity, but it seems that the head of the child was placed at the toe end of this vessel. This may be a function of vessel shape, since this amphora is long and cylindrical rather than being much wider at the high shoulder as is the case with the other two in this series in which the child had been placed with its head towards the wide end. The degree of disturbance is clearly indicated by an old break in the left femur, and the scattered positions of bits of the right radius and ulna. Two earbones were recovered from the right temporal, but none from the left. This suggests that the child may have been placed with its left ear lying down, or on its left side.

794Despite the serious disturbance of many bones, the total length of the skeleton as it appears in this container is approximately 56 cm. How well this ‘stature’ corresponds with the suggested dental age is not certain, but this finding correlates well with information summarized by Ubelaker53.

795Of general interest is the state of the mandibular symphysis, which has not yet fused. The symphysis is far from being a regular vertical surface, but is quite irregular. The lack of evident fusion, with perhaps another 4 to 6 weeks needed before fusion begins in earnest. This correlates fairly well with other data available on the rate of sympheseal function. The evidence suggests the age of this child nearly as well as does the degree of dental development, although at the lower end of the range as predicted by Becker54. Becker found fusion taking place at 6 to 9 months, based on an ancient population from Anemur, Turkey.

796The left mandible is largely intact and has in situ d2M through d1Im all in correct position to erupt, but the 1M is not evident. These teeth were not removed to determine the extent of crown and possibly root development, but the d11 is just beginning to be positioned to erupt. The lateral I is slightly smaller and lower in the jaw. The canine top is clearly formed, but it is far from ready to erupt. The d1M crown appears to be formed, and the second molar crown is most definitely formed. This pattern is the reverse for the Ml-M2 sequence as outlined by Ubelaker55.

797The right mandible has all deciduous sockets intact, and most of the socket for the Ml. Only the right maxilla is clearly represented and has three teeth in situ and a fourth (dI2) recovered from the soil. Only its canine is missing. The dM1 is still rotating, being tilted forward at about 40 degree angle. The dM2 has already achieved appropriate position for eruption. These teeth have a development similar to those of the left maxilla. The lefter alveolar area is represented only by fragments, but four of the teeth were also recovered by sifting soil. The central incisor crown is complete and also has 1.5 mm of its root developed. The lateral incisor crown has just completed development and the root is beginning to form. About 4/5 of the d1M crown is complete, but only 2/5 of the d2M crown.

798Suspected Tomb

799Age: 14-17 years.

800Five human bone elements were found in H IV which appear to derive from the right foot of a single adolescent. These are the right first metatarsals and its unfused proximal end; the right second and third metatarsals for which the distal ends have not fused, but the proximal ends have; and a terminal pedal phalange from which the unfused proximal end also has not been recovered.

801The clustering of these bones is such that they could only derive from an undisturbed burial context. However, no burial or indication of burial was noted in this area. I suspect that the burial from which these remains derive was nearly entirely destroyed in antiquity, leaving only a portion of the foot end of the grave undisrupted.

802Flecker (1942) found that metacarpal fusion was complete in females by age 14, and in males by age 17. The size of the bones in this sample suggests to me that we are dealing with a male, but this is based on very subjective observations.

803Conclusions

804The three individuals from these amphorae are all infants. The size of the individual containers are each more than adequate to contain these remains, but there seem to be some relation between the volume of the container and the size of the infant placed within.

805Formation of the second deciduous molar crown in advance of the completion of the first differs from the sequence as generally indicated by recent studies (but, cf. data from Ficana)56.