The excavation took place in five areas (A–E): Areas A–C were excavated in 2006 (Sa‘id 2008, and see background and references therein; Fig. 1: A-4930) and Areas D and E, in the current project. A burial system from the Roman period was unearthed nearby in a previous excavation (Yankelevitz 2008).
 
Area D
Two phases of construction from the Byzantine period were discovered: the earlier includes a quarry for kurkar building stones, and the later includes a building with at least four rooms (I–IV; Fig. 2). The building was constructed after the quarry went out of use, and the area was leveled with a fill of soil mixed with a large quantity of sherds. Four mosaic floors (a–d) were found in the building. Floors a, b and d were polychrome, while Floor c was white. A few sherds from the Persian period from the surface indicate activity at the site in the fifth century BCE.
 
Mosaic Floor a featured geometric patterns in black, red and white (L408; Fig. 3). The mosaic floor was demarcated by a frame of rhomboids bounded by two black bands that ran along the robber trenches of the walls that once surrounded the room. The geometric pattern within the frame featured diagonal black bands creating crosses, or, when viewed from another angle, rhomboids with scales (Type J5; Avi-Yonah 1981:142). This pattern has been identified in many churches, for example, at Horbat Hesheq (Aviam 2002:179, Fig. 19), in the Monastery of Martyrius at Sukhmata (Magen 1993: Pl. IX) and at Herodium (Netzer, Birger-Calderon and Feller 1993:222–232).
 
Mosaic Floor b. A small patch of floor that survived north of Floor a (L409; Fig. 4) featured one form of pattern belonging to Avi-Yonah’s Type H3 (Avi-Yonah 1981:287)—squares formed by an outer black frame and an inner red one, with a Type D1 pattern in red in the center of each square and a Type D3 pattern between the squares, also in red. Black lines projecting from the corners of the squares intersect to create crosses.
 
Mosaic Floor c is a white and simple undecorated monochrome mosaic (L413; tessera size c. 1 cu cm). It may have been part of a side room that was less significant in terms of representation but still larger than the other rooms.
 
Mosaic Floor d is a polychrome floor decorated with geometric and vegetal patterns (L419) of Avi-Yonah’s Type H, including Patterns H3 and H11 (Ovadiah and Ovadiah 1987:244). The northwestern end of the floor, in a band bordering Floor a, was decorated with an amphora flanked by two birds facing each other besides the amphora handles. The design is surrounded by clusters of grapes and a grapevine with leaves resembling ivy leaves (Fig. 5). Under the band of leaves is a rectangular panel enclosed in a frame of rhomboids adorned with geometric patterns. In the frame under this panel is an inscription from which four Greek letters survived: ETHХ. According to Leah Di Segni, this belongs to a type of dedicatory inscriptions that read “May so-and-so, layer/builder of the mosaic…”. In the southwestern corner of the carpet, the patterns in the frame feature three pomegranates (Fig. 6). The edges of the carpet consist of a white band surrounded, in turn, by a red strip of Pattern A5–6. The carpet is made of red, black and white stones and decorated with rhomboids interspersed with trapezes. Each of these components contains polychrome geometric designs. Similar geometric patterns have been found in a church at Magen (Tzaferis 1993:284). Two stone fragments—of a basin (Figs. 7; 10:1, below) and of a tombstone (Fig. 10:2)—were found in a probe to the south of a wall (W5) that abuts Floor d.
 
Pottery. The topsoil layer yielded a few sherds from the Persian period (fifth century BCE), including black, burnished Attic bowls (Fig. 8:1) and two jars (Fig. 8:2, 3). The Byzantine period is represented by fragments of numerous LRRW bowls (Hayes 1972; 1980), including ARS bowls (Form 61—Fig. 8:4, 5; Form 58B—Fig. 8:6), dated to the fourth–fifth centuries CE, many CRS bowls (Form 9B—Fig. 8:7, 8; Form 10—Fig. 8:9; Form 11—Fig. 8:10; Form 7—Fig. 8:11), dated to the late sixth–late seventh centuries CE, as well as PRS bowls (Form 3—Fig. 9:12; Form 2—Fig. 8:13; Form 5—Fig. 8:14), dated to the fifth–sixth centuries CE and mortaria from the northern Syrian coast (Fig. 8:15, 16), dated to the second–fourth centuries CE. Cooking pots were also found (e.g., Fig. 9:1), many bag-shaped jars (e.g., Fig. 9:2), Gaza jars (Fig. 9:3, 4) and imported amphorae (Figs. 9:5–8) from North Africa, Cyprus and the northeastern Mediterranean, some of which were intended for maritime transport. Also found were a fragment of a slipped mold-made flask originating in Sagalassos in Asia Minor and dating from the sixth century CE (Fig. 9:9) and a disk lamp (Fig. 9:10), perhaps of the Caesarea type, decorated with the figure of an animal (for riding?), a bell attached to its neck and a saddle or saddlebag on its back. The lamp was dated to the mid-fourth–fifth centuries CE.
 
Stone items. The abovementioned basin is semicircular and made of kurkar stone (inner depth 0.29 m, outer height 0.52 m, outer diam. 0.72 m, inner diam. 0.48 m, width of open side 0.68 m; Fig. 10:1). There is an opening in its oblate side, and a groove is hewn into its base. Remains of plaster were preserved on its inside and outside. A similar basin was found in a church at Shivta (Tsafrir 1984:231), where it was adjacent to a large, deep baptismal fonts for adults. The small basin was intended for baptizing babies or perhaps to hold oils with which to anoint the baptized. The probe in which the basin was found yielded also a marble slab with a bilingual inscription in Latin and Greek reading “C SITUS”, translated by L. Di Segni as “Here lies”—typical of tombstone inscriptions (Fig. 10:2). North of Floor c, a fragment of a marble chancel screen panel decorated with a cross was found (Fig. 10:3). Another fragment of a marble chancel screen panel was found on Floor a (not drawn). Under this floor was a third fragment of a marble chancel panel decorated with a cross (Fig. 10:4) along with cut stone items, perhaps inlays (Fig. 10:5, 6), similar to an object decorated with a cross found at Kh. ed-Deir in the Judean Desert (Hirschfeld 1993:246).
 
Metal objects. Iron nails were found that are square in section and have a broad, dome-shaped head (not drawn).
 
Area E
Three superimposed layers were found. The earliest was a layer of stream pebbles, apparently naturally deposited when they were swept down from the slopes of the Carmel, which rises steeply above the excavation area. Above the layer of pebbles were two clearly differentiated construction layers (Fig. 11), both from the Byzantine period. The early layer includes two well-plastered pools, the southern of which was built of ashlars and the northern of debesh. It seems that they were intended for the impounding or treatment of liquids; they may have been used to raise sea fish until ready for consumption. The later layer included remains of structures, some overbuilding the walls of the earlier pools and some on the natural ground. The wall foundations were built of fieldstones. Of the structure, only the foundations were preserved, below the level of the floors.
 
Pottery. LRRW bowls were found, including some types of the ARS family (Form 104C—Fig. 12:1), dated from the mid-sixth–seventh centuries CE, CRS bowls (Form 7—Fig. 12:2; Form 2—Fig. 12:3–6; Form 4—Fig. 12:7; Form 9B—Fig. 12:8), PRS bowls (Form 3—Fig. 12:9–18), dated from the fifth–seventh centuries CE and a bowl from the FBW family (Fig. 12:19; Magness 1993: Form 1A), dated from the sixth–early eighth centuries CE. Also found were mortaria from the northern Syrian coast (Fig. 12:20, 21), dating from the fifth–seventh centuries CE, cooking pots (Fig. 13:1, 2), some possibly from Cyprus, bag-shaped jars (e.g., Fig. 13:3), imported amphorae (Fig. 13:4–7) and a pithos (Fig. 13:8) from the northern Syrian coast, dating from the fifth–seventh centuries CE. A bowl from the Ottoman period (Fig. 13:9) probably represents the phase of the construction of the Carmelite and St. Joseph monasteries in the complex in the nineteenth century CE.
 
Glass vessels. The glassware includes fragments of stem lamps as well as bases of wine goblets. Fragments of plates, juglets and glass bottles (not drawn) dated to the Byzantine period (sixth–seventh centuries CE) were also found.
 
Stone objects. A fragment of a marble altar table was found (Fig. 8:7).
 
Metal objects. Numerous iron ingots were found, attesting to the presence of an ironsmith workshop at the site. A cylindrical, elongated and folded lead weight was also found, of a type attached to the bottom of a fishing net (Subtype L2.3; Galili, Rosen and Sharvit 2010:90).
 
Coins
Robert Kool
 
Fourteen coins were identified in the two excavation areas (Table 1), all from the Byzantine period (324–604 CE). One of these, found on Floor 408 (Mosaic Floor a), is a rare coin of Emperor Phocas from the Ravenna mint. Minted in 604 CE, this is the latest these coin, apparently dating the final days of the building, which was probably destroyed during the Sasanian conquest in 614 CE.
 
Table 1. The coins from the excavation
IAA No.
Locus
Basket
Date (CE)
Ruler and Mint
138857
420
96
324–497
 
138860
423
109
324–497
 
138852
408
13
500–620
 
138856
420
92
500–599
 
138859
423
104
500–620
 
138853
408
36
507–512
Anastasius I, Constantinople
138854
418
79
518–522
Justinian I, Constantinople
138858
412
47
518–522
Justinian I, Constantinople
138862
511
215
518–522
Justinian I, Constantinople
138863
511
244
518–522
Justinian I, Constantinople
138864
408
-
518–522
Justinian I, Constantinople
138855
418
80
538
Justinian I, Constantinople
138861
507
214
539
Justinian I, Constantinople
138851
408
12
604
Phocas, Ravenna
 
Conclusion
The site was first settled in the Persian period, apparently around the fifth century BCE. During the Byzantine period a structure with polychrome mosaics stood at the site. The mosaic is dated to the sixth century CE, and judging by the pottery assemblage it can be concluded that the structure dates either to the sixth century CE or to the early seventh century CE at the latest. The structure seems to have been either a church or monastery; two modern-day monasteries at the site attest to the sanctity of this location. It appears that the site was abandoned in the early seventh century CE, probably with the Sasanian conquest in 614 CE.
The excavation findings call into question the identification of Haifa of the Mishnaic and Talmudic periods with the site of Haifa el-‘Atiqa. An examination of the historical sources along with the archaeological findings does not support this identification. The relocation of a settlement was not a rare event; suffice it to recall that Haifa was moved over the past 300 years by Daher el-‘Omar, the Bedouin ruler of the Galilee, who relocated the city about 2 km eastward—from the site identified as Haifa el-‘Atiqa to the slope extending from today’s Hadar Ha-Carmel down to the Lower City—severing its connection to its previous location.