Mersa/Wadi Gawasis 2006-2007 Report, Rodolfo Fattovich and Kathryn A. Bard editors

III.2. Archaeology - Artifacts

Claire Calcagno, S. Terry Childs, Giulio Lucarini, Mohammed Mustafa Abdel Maguid, Andrea Manzo, Cinzia Perlingieri, André J. Veldmeijer, Chiara Zazzaro

Pottery (Cinzia Perlingieri)
Most of the ceramics from all 2006–07 excavation and stratigraphic units are Middle Kingdom in date, from the early 12th to late 12th-13th Dynasties.
Pottery from excavation units WG 37/38/42 and all geological test pits, consists mainly of large fragments of early 12th Dynasty bag-shaped storage jars, Marl C and rare Marl A3 wares. Pottery from these units is exceptionally well preserved. A few later collared jars with flat rims, dating to the late 12th and 13th Dynasties, were also found in these units (cfr. Bader 2001: 160-163).
Pottery from the excavation units at the upper western slope (WG 32, WG 33, WG 40) and from the interior of Cave 3 (WG 39) is consistent with a dating of the late 12th and 13th Dynasties. Marl C ovoid/globular jars with corrugated neck are common; they can be also found during the Second Intermediate Period, but there is very scarce evidence for such a late chronology (see Bourriau 2004: 8-11; Do. Arnold 1982: 11:7). Marl C collared large storage jars, with a direct rim with a flat termination, may also date to the late 12th–13th Dynasties based on parallels with the ceramics from Tell el-Dab'a (see Bader 2001: 160-163). Nile Silt small unrestricted hemispherical "drinking" bowls and small- to medium-sized plates are common in these assemblages.
A large assemblage of Marl C, with some Marl A3, medium-sized round bottomed plates were found in WG 32, SU31, piled at the entrance of Cave 6 (Figure 13 a ). Medium- and large-sized jars were also found in the same assemblage. Thirty-three complete plates and many large sherds of plates were excavated. The plates have a diameter of ca. 16-19 cm. Many of them carry pot marks (single or multiple commas), which were engraved before firing. The outside of the bottom of all plates was hand-finished. This type of plate is well documented for the entire Middle Kingdom and later. Similar plates were found in different deposits at the south pyramid and the cemetery at Lisht, where they have been interpreted as jar stoppers because of the occurrence of lumps of mud used to seal the jars still adhering to some plates (Do. Arnold 1988: 110). Unfortunately, the plates from WG 32 were covered with a crust of salt, and no trace of sealing mud was preserved.
The pottery from the production area (WG 19/25/26/27/44) confirms the two main chronological phases of use, identified in previous field seasons. The earlier phase of use dates to the early 12th Dynasty, as can be inferred from the occurrence mainly of large restricted necked jars ("bag-shaped" zirs), which are very common in the early Middle Kingdom (Do. Arnold 1988; Bader 2001: 160-163; Bader 2002: 42-43). The later phase of use dates to the late 12th –13th Dynasties. This dating is based mainly on the massive presence of ovoid/globular jars with corrugated neck (a type common during the late 12th and 13th Dynasties).
A complete Nile B2 carinated necked jug with thickened everted rim and conical base was found nearby in a possible fire pit (WG 26, D4-C4, SU105), lying on the side. The vessel is 25 cm. high with a maximum diameter of ca. 11 cm. This shape occurs in Middle Kingdom assemblages, though it is not a very common type. Similar jugs were found at Lahun (Kelley 1976: Pl. 40.5, 13P, 13S), Kahun (Kelley 1976: Pl. 40.12, 73), and Harageh (Kelley 1976: Pl. 44.3, 13M
[11] ).
A typological/functional analysis of the forms is still in progress. At present, we can state that the pottery assemblages from WG 19/25/26/27/44 are characterized by utilitarian coarse ware, with a high percent of locally made pottery. Conical bread molds, platters/ovens, large plates and large- and medium-sized jars in Nile B2 and Nile C wares are common. Pottery from units WG 37/38/42 and the geological pits consists mainly of storage vessels made of wares from the Nile Valley. Pottery types from all phases on the upper western slope (WG 32, WG 33, WG 40), mainly jars, bottles, drinking bowls and small to medium-sized plates, in association with other evidence for administrative activities (sealings, clay sealing fragments, cargo boxes), seems to suggest that the area outside the caves was largely used for organizational and administrative purposes.

Exotic ceramics (Andrea Manzo)
In the 2006–07 field season 8 exotic potsherds were collected at Mersa/Wadi Gawasis, but only 3 fragments could be classified and compared (Figure 13 b ):
1) A grey ware body-sherd, with dark grey polished internal surface, reddish-brown external surface, and medium-sized mineral inclusions, from WG 39 A10, SU11. Incised and impressed decoration on the external surface consists of parallel grooves covered with comb impressions. Similar decorations were recorded in the Eritrean-Sudanese lowlands at Agordat (Arkell 1954: 58-59, Fig. 27, Pl. VI, 4). This potsherd is also similar to the Gash Group Basket Ware of the early to mid-2nd millennium BC.
2) A grey ware body-sherd, with smoothed grey internal surface, grey external surface, and small-sized mineral inclusions, from WG 33, SU1. The external surface has an incised criss-cross pattern within impressed wavy-patterned bands produced by making triangular impressions. The decoration is similar to some Kerma potsherds from sites in the Fourth Cataract region dating to the first half of the 2nd millennium BC (Kołosowska, Mahmoud el-Tayeb and Paner 2003: Pl. 6; Wolf 2004: Pl. 4), and some potsherds from Agordat (Arkell 1954: Pl. VI, 6; Pl. VII, 4; Pl. VIII, 5).
3) A brown ware, mineral tempered potsherd characterized by roughly smoothed surfaces and two deep and large parallel grooves on the external surface. It may be similar to some decorative patterns of South Arabian Tihama ceramics, [12] but the identification needs further confirmation.
Two potsherds excavated in 2006-07 have no definite similarities to known ceramics from the surrounding regions:
1) A fragment of a closed vessel with a slightly everted and thickened flat rim, mineral tempered brown ware, with smooth surfaces and a band of notches parallel to the rim on the shoulder, from WG 19 A5, SU106.
2) A fragment of a close vessel with a slightly everted rim, mineral tempered brown ware, and smooth surfaces from WG 26 C4 SU97.
Two other potsherds, from WG 33, SU2, are body fragments of an undecorated hand-made ware, whose origin cannot be identified by a macroscopic investigation of the paste. Another potsherd, from WG 38 A3/A 4, SU7, is characterized by traces of incised lines, but unfortunately no comparisons can be suggested because of its small size and bad preservation of the surfaces.
Although small in number, exotic potsherds collected in the 2006–07 field season provide the first evidence at Mersa/Wadi Gawasis of possible contacts with the western Eritrean-Sudanese lowlands to the southeast of Nubia, and may support the hypothesis that Punt possibly included the coastal and inland regions of both sides of the southern Red Sea (e.g. Fattovich 1996; Manzo 1999).

Wooden boxes [13] (Andrea Manzo)
Twenty-two complete or fragmentary wooden boxes were excavated in WG 32 in the same Middle Kingdom assemblage where other 21 similar boxes were found in 2005–06. They had been piled altogether and were from the same shipment. The boxes were covered and filled by a deposit of windblown sand and sea grass. Some of them had been damaged by termite activity. An inscription on a box discovered in 2005-06 suggested that they were used to transport products from Punt (Fattovich and Bard 2006).
Nine boxes measured ca. 50-52 cm x 32-34 cm, like those discovered in 2005–06, and in some cases were 24-27 cm high. Four boxes measured ca. 45-48 cm x 30-34 cm and 20 cm high. Four boxes measured ca. 51-55 cm x 37-39 cm high. The planks were always ca. 2 cm thick. Five boxes were very fragmentary and only the thickness of the planks was recorded.
The sides, bases, and lids of the boxes were formed by two or more planks, matched and connected by dowels. The external and internal surfaces of the boxes were usually covered with plaster. A light red color is often visible beneath the plaster coating. Frequently, a thick layer of white plaster, different from the plaster of the coating, was found on plank joints, and, in the case of Box 41, formed a stratum in the box fill. This plaster was probably intended to better seal or preserve the content of the boxes.
All boxes are similar in typology and construction technique. The joints between planks were fixed with dowels with a round cross-section, 2.5-10 cm long and 0.5-2.5 cm in diameter. The base, when preserved, always had rectangular furring strips. On one lid and on a short side of a box conical pegs were preserved. They were intended to close the boxes with ropes. Their occurrence confirms that most of the clay sealings with impressions of a peg, rope, and wood grain were originally used to seal the boxes.
Box 21 had a hieroglyphic inscription painted in black on the external surface of one of the short sides confirming that these boxes were used to transport products from Punt. Shorter inscriptions consisting of few signs were found on Boxes 36 (possibly a fabrication mark), 40, and 41.
Boxes similar to those from Mersa/Wadi Gawasis were also found at Kahun in an assemblage dating to the 12th Dynasty (David 1986: pl. 4). [14]

Ship components [15] (Claire Calcagno and Chiara Zazzaro)
Significant amounts of wood material with evident nautical associations were recovered from Mersa/Wadi Gawasis as in previous field seasons, including recycled ship timbers, joint elements, and wood debris generated as a by-product of timber reworking.

Timbers
Twenty timbers were identified as ship components during the 2006–07 field season.
Ship timbers examined microscopically by Rainer Gerisch were identified as acacia from the Nile Valley (Acacia nilotica) and imported cedar (Cedrus lebani). The condition of the wood ranged from soft, spongy and crumbling, to hard and salt-encrusted or burnt. Some pieces survived only as traces in the sand; several were too decayed for recovery. A number of timbers are considerably reworked, such that their original nautical function is difficult to determine.
The ship timbers had been recycled at the site in various ways. Four timbers were recovered outside the cave entrances. Three appear to have been reutilized in walkways to the cave shelters, including T19 and T58 outside Cave 2, and T53 within the entrance to Cave 6 (WG 32). Timber T71 was found loose in the accumulated sand deposit outside of what may have been the entrance to Cave 4 (WG 33). One timber (T51) was located in Cave 2, Room 1, in a context recently disturbed by thieves. It was most likely associated with a previously studied assemblage of four timbers from that area which formed a walkway (see Zazzaro in Bard and Fattovich 2007).
Cave 3 (WG 39) contained 14 timbers (see 'WG 39 Excavation Report' above). The orthogonal disposition of six of these suggests a deliberate arrangement. Several fragmentary timbers were also found loose in the windblown sand deposits, or associated with hearths. Four timbers (T61, T64, T66, T67) remain partly wedged beneath collapsed cave rock and could not be examined in their entirety this season.
Eight timbers have been provisionally interpreted as hull planks based on the presence of shipworm damage and/or their fastening patterns. These include two small thick timbers beveled at one end (T62, T63), which M. Abdel Maguid suggests could have functioned as stealers.
Timber T64 is a thick, heavily reworked timber that is also likely to have originally functioned as a hull plank. It is 87 cm long, at least 46 cm wide and 22 cm thick. It features gribble on one face, a complex variety of joints (see below), and a possible scarf cut. Blocked beneath a cave collapse, it remains to be fully excavated.
Two timbers (T51, T65) with beveled ends on one face might correspond to deck planks. However, T65 (measuring 86 cm x 27 cm x 6 cm) might have originally functioned as a hull plank: it has been intensely reworked, features a mortise-and-tenon fastening, and contains large knots.
While timbers T66 and T67 have not yet been fully excavated, they could be identified as deck beams as both feature a longitudinal raised central section on one face. Timber T66 can be compared to the deck beam identified last year (T32; see Ward in Bard and Fattovich 2007).
Three poorly preserved but distinctive timbers (T55, T68, T70) may not belong to the hull structure itself based on their particular shapes and lack of gribble or typical hull timber fastenings. The largest (T55) is curved and features a square hole and peg toward one end; it extends over 370 cm in length (although only ca. 130 cm was found in structurally sound condition), and measures 20 cm in width and 40 cm in thickness (Figure 14 a ). Timber T68 is a short round-sectioned pole tapered at both ends. Timber T70 is pronged at one end and has a flat rectangular section. These timbers could be related to a vessel's superstructure, or otherwise have another function related to maritime activities at the site.
Four timbers (T54, T56, T58, T69) are in extremely poor condition, and their identification remains problematic.
Several forms of fastening were recorded this season, including single and paired mortise-and-tenon joints, dovetail mortises, pegs, and one lashing channel, in addition to dozens of loose tenons.
Timber T64 is remarkable for its variety of joints (Figure 14 b ). On one edge it features two pairs of double mortise-and-tenon joints that average 10 cm in width and 2 cm in thickness. A through mortise containing traces of copper is carved through the plank thickness, and connects to a lashing channel that penetrates diagonally into the through mortise from the plank edge. Two pegs on the plank's visible wide face might also be associated with edge fastenings, although their function remains to be clarified following complete excavation of the timber. Especially noteworthy are two dovetail mortises cut into opposite sides of the wide face. They measure up to 14.5 cm in length, between 3.4 and 6.5 cm in width, and up to 3.8 cm in depth; thus they correspond in dimensions to the range of loose dovetail tenon halves recovered in Cave 3. Significantly, these mortises constitute the first indisputable evidence of dovetail joints utilized in an Egyptian ship timber in ancient times.
Like T64, timber T65 is also mortised through its thickness. In both cases, the mortise contains traces of copper which are likely to be remnants of metal ligatures such as those first noted in the 2005-06 field season (e.g. T34; see Ward in Bard and Fattovich 2007). It is hypothesized that these copper 'staples' may have been used for their antifouling properties to protect the fastening exposed on the immersed side of the hull.

Wood debris
Wood remains that could not be clearly identified as ship timber parts included joint elements (loose tongue tenons and dovetail tenon halves), fragments of worked wood, and collections of wood debris (lots W167-W377). Wood debris and joint elements were recovered in quantity in trench WG 39 (Cave 3), as well as in trenches WG 32, WG 33, WG 40, WG 41 and WG 42. Most of these materials are likely related to ship dismantling and careening processes (see Ward in Bard and Fattovich 2007).
The greatest concentrations of wood debris (almost two-thirds of the total collected) occurred in Cave 3 (surface collections, SU1, and SU11). Wood species identified among samples of wood debris in this trench included acacia, sycomore and cedar (see Gerisch, this report). Noteworthy is the recovery from within the cave of several pieces of oak (Quercus sp.): a wood fragment (W376), and an assemblage of related charcoal pieces (A9, SU12) (ibid., this report), which may be related to other nautical material found in the area.
The debris includes a number of pieces with significant traces of gribble, attached barnacles, and evidence of tool marks and surface stripping. A small number of wood fragments bear traces of what appears to be copper (e.g. W229, W280), while others associated with hearths recorded in Cave 3 show evident signs of burning.

Tenons
A total of 52 loose tongue tenons varying in size, condition and features were recovered, primarily in Cave 3 and WG 32. Intact tenons range in length between 8 cm and 27 cm, and in width between 4.2 cm and 8.5 cm; thicknesses generally range between 1 cm and 1.5 cm. Six tenons feature pegs and peg holes (or partial holes) which vary in diameter between 1 cm and 1.5 cm. All tenons analysed are of Acacia nilotica wood (see Gerisch, this report).

Dovetail tenons
Twelve dovetail tenon halves were recovered this field season: 10 from WG 39, 1 from WG 24, and 1 from WG 32. Most show adze marks where they were broken in half at their narrowest point. Dimensions generally range between 11 cm and 18.5 cm in length, between 4.4 cm and 6 cm in width, and between 2.8 cm and 4.8 cm in thickness. Analyzed dovetails are of Acacia nilotica (see Gerisch, this report).

Other
Among the remaining wood fragments are several with distinctive features. These include five similarly fragmentary round-sectioned pieces whose estimated original diameters range between 4.5 cm and 8 cm (although none survives with more than ca. 60% of its original diameter).

Anchors (Chiara Zazzaro and Mohammed Mustafa Abdel Maguid)
During the 2006-07 field season at Mersa/Wadi Gawasis five new stone anchors, two complete and three fragmentary, were discovered. Some fragments of limestone slabs, possibly pieces of anchors, showing chisel marks on the surface, were also recorded. The stone anchors were studied and numbered from A21 to A25, continuing the inventory numbers of the Gawasis general catalog of anchors (Bard and Fattovich in press). It was also possible to better record the dimensions of anchor A1, a granite anchor recorded in 2003-04, because thieves, attempting to enter Cave 3 after the 2005-06 field season, displaced it from its original position.
Geologist Mohamed Badr selected 14 stone samples for petrographic analyses from anchors at the site. Limestone used to make anchors will be examined and compared with Egyptian sources and/or ancient quarries. The goals are to better understand the production process of the Mersa/Wadi Gawasis anchors and, consequently, to investigate an important part of the organization and provisioning of seafaring expeditions.
All identified stone anchors were described, photographed, and drawn to scale, recording dimensions and, if possible, also the weight. Anchors recorded during this field season have the typical rounded top, with a hole and groove for rope. Two complete anchors (A23 and A24) represent a new typology (Figure 15 ) because they are rectangular in shape and smaller (21-24 cm in length) than anchors previously recorded at Mersa/Wadi Gawasis. A24, found in the lagoon bottom as was the larger one A11 (2005-06 field season, see Fattovich and Bard 2007), demonstrates that ships approached the lagoon with anchors of different dimensions. Possibly small boats that could easily enter the lagoon used such small anchors, or small anchors were used together in a chain (Wachsmann 1997: 286-287).
A23 was found reused for construction in the northwest wall of Cave 2 and had been placed at the entrance between the cave's ceiling and a timber (T20). This is the seventh anchor recognized in the Cave 2 walls, after the six larger ones discovered during the 2004-05 field season (Bard, Fattovich et al. 2005).
The fragmentary anchors A21, A22 and A25 were found in WG 33, SU2, near the Cave 4 entrance. They had possibly been reused in the cave structure. Comparing these partially preserved anchors with the dimensions of the complete anchors already discovered at the site, it is possible to estimate their original dimensions. The fragmentary small anchor A25 is comparable, in the estimated dimensions (possibly 40 cm x 22-24 cm), to the small anchor found by Abdel Moneim Sayed during his excavations at Wadi Gawasis on the eastern terrace (Sayed 1980: XXII: 3)[16].
The shape of the top of anchor A21 is similar to the complete limestone anchors found in the walls of Cave 2 during the 2004-05 field season. The type of groove on A21 and its dimensions are similar to those of A4 (see Bard, Fattovich et al. 2005). The estimated length of A21 is ca. 98-100 cm.

Ropes (André J. Veldmeijer)
One of the caves at Mersa/Wadi Gawasis, Cave 5, contained large amounts of ropes from Middle Kingdom times. The ropes are coiled longitudinally and wound in the middle as to allow easy storage for future use. This technique of storing long ropes is still used nowadays to prevent it from getting entangled. The back of the cave contained the majority of the rope coils; toward the front the number decreased rapidly. Remarkable is the fact that there are two smaller coils situated on top of the windblown sand that now obscures the original entrance to the cave. Further back in the cave, the visible layer of coils is the top layer; at the back of the cave there is, at least partially, a second layer.
The preservation is remarkable, but the ropes are very brittle. Many coils have fallen apart and only a few are more or less complete. The small holes in the ropes are remnants of insects, but it has yet to be determined by what kind of animal.
All ropes are made of the same material, macroscopically identified as halfa grass. The composition is sZ3, which is an often-used composition. The ancient Egyptians used a third yarn and increasing diameter rather than cabling to create stronger ropes. The diameters of the various coils varied from about 24 mm to as much as almost 40 mm for the ply. The estimated length of the most complete coil seems to be 30 m. A detailed analysis of these diameters and the Cord Index of Ply is in progress.
In the top layer, 16 more or less complete coils have been identified along with two big piles of rope, possibly consisting of three or more coils. These have been drawn on a map. It is clear that not all rope was deposited as coils, or, more likely, not as comparable coils. The second layer, probably extending in half a circle from the collapsed wall, which now serves as the entrance, might contain an additional 10 coils, bringing the estimate of the total number of coils to about 30.
Next season the work will focus on completing the description of the find as well as weighing coils.

Stone assemblage (Giulio Lucarini)
General remarks
The entire stone assemblage collected at Mersa/Wadi Gawasis during the 2006-07 field season includes 1017 chert artifacts, 16 so-called large stone tools (saddle querns, upper grinders, rubbers and large stones of undetermined function), and 13 quartz fragments. The 1017 chert artifacts include cores (25: 2.5 %), debitage elements (983: 96.6 %) and retouched tools (10: 0.9 %).
The artifacts were collected in excavation units WG 19, WG 24, WG 25, WG 26, WG 38, WG 32, WG 33, WG 39, WG 40, WG 42 and WG 44. Only two pieces were collected on the surface.
The majority of the stone implements were from the activity area in WG 19/25/26/27/44 at the base of the western slope. Very high concentrations of lithic artifacts were recorded in WG 26 (543 lithics and 8 large stone tools) and WG 44 (380 lithics and 2 large stone tools).
The other areas of the site are characterized by a lower number of artifacts. Four lithic artifacts and 2 undetermined large stone tools were found in WG 32 and WG 33, in front of the cave entrances. An upper grinding stone was found inside Cave 2 (WG 24).
Finally, 34 lithic artifacts and 1 lower grinding stone were collected in the trenches at the base of southern slope (WG 38/42, WG 39, WG 40). Considering the very low number of artifacts from these areas, the frequency of cores, debitage elements, and retouched tools was analyzed only on the artifacts from WG 19/25/26/27/44. The raw material color was classified with Munsell Soil Color Charts (1990 Revised Edition).

Main techno-typological traits of the lithic assemblage
Raw material: chert was the only raw material used in manufacturing the lithic artifacts, which were recorded in the 2006–07 field season. The few quartz fragments do not seem to have been intentionally knapped. The chert is generally characterized by a fine texture and high homogeneity even if pieces with several inclusions occur as well.
From the morphology and size of the implements it is likely that medium and large size chert pebbles, available along the bed of the wadi, were mainly exploited. Almost all the artifacts are well preserved, even if some of them are fragmented. A few pieces with a reddish (5YR 5/2) or dark reddish (5YR 4/2) color are burned.
Cores: the 25 cores represent the 2.5% of the Mersa/Wadi Gawasis lithic assemblage, ranging from a lowest frequency in WG 44 (1.8%) to the higher one in WG 19 (4.3%) and WG 26 (2.6%). Most cores, with a small size and sometimes microlithic, usually show a high or medium degree of exploitation. The multiple platform type is the most frequent (11 cores: 44%), followed by the single platform type (8 items: 32%) (Figure 16 a ) and 90° platform cores (5 items: 20%). An opposed platform core found in WG 26.
The striking platforms often do not show any trace of preparation. Only a microlithic, single platform core from WG 44 shows tiny edge strengthening removals on its striking platform. The negatives of removals on flaking surfaces provide clear evidence of an almost exclusive production of flakes.
Debitage: debitage elements (983 artifacts: 96.6%), are the most represented category in all excavation units of the site, ranging from 96.7% in WG 26 to 97.1% in WG 44. In these two trenches, the high frequency of debitage elements may be partially related to the high number of chips and chunks; 327 and 232 debris elements have been collected in these two units, respectively (ca. 62% in both trenches).
Flakes are the largest class of debitage elements (296 specimens, with a frequency ranging from 27.1% in WG 44 to 30.9% in WG 26). Flakes are mainly from single platform cores (141 specimens), with a frequency ranging from 40.7% (66 pieces) in WG 26 to 51 % (51 pieces) in WG 44. Primary flakes and early stage core working elements are the second well represented class (65 artifacts: 21.9%). Multiple platform flakes are also well represented (57 elements: 19.3%. 90° platform flakes (20 specimens: 6.8%) and opposed platform flakes (4: 1.4%) are less frequent. It was not possible to determine the type of the remaining 9 flakes.
The blade index is quite low (22 blades: 2.2%). The highest number of blades was found in WG 26, where the 10 pieces are the 1.9% of the debitage products found in the trench. Nine blades come from WG 44 (2.4%). Almost all specimens are from single platform cores (18 pieces: 81.8%). Two multiple platform blades, an opposed platform blade, and an unidentifiable type are present as well.
Core rejuvenation practice is mainly documented by 62 core trimming elements (6.3%). They are particularly frequent in WG 26 (26: 4.9%) and WG 44 (28: 7.6%).
Retouched tools: the 10 retouched tools show the lowest frequency (ca. 0.9%) in the whole lithic assemblage. They include 3 side-scrapers, 3 notches, 2 denticulates, 1 perforator, and 1 piece with a continuous retouch.
Two side-scrapers from WG 26 are manufactured on core sides, with an inverse, flat retouch (Figure 16 b ) and a bifacial, semi-abrupt retouch, respectively, and one from WG 32 on a primary flake with bifacial, flat detachments. All three artifacts seem show a very smoothed retouched edge suggesting a long time of use.
The two denticulates from WG 39 and WG 44 show a primary flake and a core side as blanks, respectively. They are characterized by an obverse, slightly invasive, semi-abrupt retouch.
The notch from WG 26 is manufactured on a core side (Figure 16 c ). The other two notches from WG 44 are on flakes. Two notches are of a "single blow" type; the third one is retouched.
The piece with a continuous retouch from WG 44 is manufactured on a tertiary flake from a single platform core, and is characterized by a rectilinear, marginal, flat retouch.
A perforator/side-scraper on a secondary flake from a single platform core was also found in the upper deposit of WG 26 (Figure 16 d ). This piece shows an inverse, denticulated, slightly invasive, retouch along the whole perimeter, with bifacial retouch on the distal end for the perforator. Multifunctional retouched tools were already excavated during previous field seasons. In particular, several perforator/side-scrapers were found on the western slope of the terrace.

Conclusions
The large amount of lithic artifacts from the base of the western slope of the coral terrace (WG 19/25/26/27/44) supports the information from previous field seasons about the massive exploitation of chert in the so-called "production area."
Observation of the lithic assemblage showed an almost exclusive presence of debitage elements, in which the flakes are the most represented class. Blades, on the contrary, occur in a lower quantity. The fairly high occurrence of core trimming elements, together with the large amount of debris, provide a clear indication of a prolonged exploitation practice of the cores, pointing to local manufacturing activities. This seems to be also confirmed by several hammer-stones from WG 26.
The tool kit does not show a particular standardization of the products, and, on the whole, the very scanty retouched elements do not provide clear typological traits of the assemblage. Pieces to be used for scraping seem to be more frequent, representing the majority of the retouched tools. Among them, the presence of some opportunistic tools (such as different kinds of rough scrapers and denticulates), could be connected with wood exploitation.
Hopefully, use-wear analysis of some artifacts will provide more information about the function of these tools. Thus, some lithic artifacts have been selected and a cast of their working edges has been produced using a hydrophilic vinypolisiloxane impression resin. The casts will be analyzed with a metallographic microscope in order to detect the possible presence of micro polishes.

Copper-based materials (S. Terry Childs)
The majority of the copper-based fragments found during the 2006-07 field season was small (less than one centimeter in diameter) and heavily corroded. They were unrecognizable as to shape and function. There also were fewer copper-based pieces found this field season than in 2005-06. Less than ten were large and intact enough to measure, describe, and speculate about function, given their context.
Similar to the fragments found in 2005-06, the majority of the sizeable, copper-based pieces were flat strips or bands with standardized widths of either 1.5 or 2.0 cm. Several band fragments from this field season were also 1.0 or 2.5 cm in width, which suggests that there were a number of standardized widths of the bands.
Most typically, the bands were bent and flattened at one end or bent in the middle during use. Several were cut at the ends. Several others were bent, folded over, and twisted in various places as though they were crumpled during removal, possibly from ship planks as they were repaired after a voyage.
The most interesting artifact this field season was found in the surface level of Cave 3, WG 39. It was an approximately 10 cm long band, 20 mm wide, that was folded over on itself three times and flattened so the entire thickness was approximately 6 mm. This carefully folded artifact strongly suggests that it was a portion of an original "reel" of copper banding that was probably brought to Mersa/Wadi Gawasis for purposes of ship building or repair.
The only other copper-based artifacts found that were not bands or flat, nondescript pieces were short rods or pins from WG 32. One was a rod or pin that was pointed at one end and either cut or broken at the other end. Excavated in SU31 of square B5, it was 3 mm in diameter and approximately 5.5 cm long. Nearby, but in SU10 of square A5, was a 9 cm long rod, approximately 3 mm in diameter, with no visible point at either end.
The larger copper artifacts should be X-rayed to determine if they have any real metal left or if the metal is all corroded. If any non-corroded metal exists, chemical analysis should be done to determine whether or not they are a bronze alloy. Metallography might also be done to determine how they were manufactured and to gain clues as to how they were used.