Microwear analysis of small recycled flakes and recycling products from the Ein-Zippori site, Lower Galilee, Israel

A microwear analysis of recycled lithic artefacts from late Pottery Neolithic Wadi Rabah and Early Bronze Age layers at Ein-Zippori, Israel included cores-on-flakes (COFs) which are discarded blanks made into cores, and the flakes detached from them. COFs may have microwear traces that formed before they were recycled. The focus here is on how blanks removed from recycled COFs were used. Discarded flakes were not used as cores to produce small blanks at Ein-Zippori because lithic raw material was scarce, but were COFs recycled so that small tools could be produced for specific tasks? Visible wear traces were present on 19 of 44 blanks produced from COFs. Microwear traces were similar to use wear Lemorini et al. (2015) observed on much older Lower Paleolithic recycled flakes from Qesem Cave, Israel. Most flakes struck from COFs had been used to cut and scrape meat and fresh hide (42%, n=8), but four were used to work wood (21%) and four others were used to cut, scrape, or whittle bone and wood (21%), and two were used for butchering and wood working (11%). One flake only had generic weak microwear traces (5%). These were expedient flake tools, made and used in an ad hoc fashion. Specific blanks do not seem to have been used for distinct tasks.


Introduction
Results of a microwear analysis of a sample of 69 recycled flaked stone artifacts from the large late prehistoric Ein-Zippori site located 2 km west of Nazareth in the Lower Galilee, Israel are presented. Recycling is modifying previously used artefacts so they can be used in other activities (Vaquero et al. 2012). Lithic artefact recycling is not the same as resharpening, since recycled tools are not repaired so they can continue to be used for the same tasks. Recycling previously discarded artefacts has been practiced since ancient times. Sometimes utilized tools are curated or cached for future use at sites where they can be recycled. In North America, both highly mobile foragers and less mobile logistical hunter-gatherers often cached stone tools that they manufactured elsewhere at sites where lithic raw material was not abundant (Yerkes and Koldehoff 2018:80). Many studies of recycled lithic artefacts in the Old World focused on the activities of mobile foragers during the Paleolithic at locations were lithic raw material was scarce, but also in places where flint was abundant (see Parush et al. 2018 and references therein). The temporal context of this functional study of flake blanks produced from recycled flakes used as cores is much later than the Paleolithic. It is during the Late Pottery Neolithic period and Early Bronze Age, when farmers in the southern Levant made and used new (and different) types of flint tools. Ein-Zippori was an example of expedient recycling by settled agriculturalists at a location where lithic raw material was abundant (Parush et al. 2018). Small flake blanks were produced quickly for tasks that needed to be completed immediately. The goals of this microwear study are to determine the function of these small flakes, identify the tasks they were used for, and learn if different types of blanks were used for specific tasks.
The microwear sample included discarded recycled flakes called cores-on-flakes (COFs) and the different types of blanks that were struck from them. Since COFs may have microwear traces on their edges that formed when they were used before they were recycled, those prior activities are not the focus of this study. In this study, the function of COFs was to serve as cores for the production of new blanks. Functions of the small blanks were determined by examining microscopic wear traces on their edges. The microwear analysis was conducted in 2014 and 2015 at Tel Aviv University and at Ohio State University.
Excavations at Ein-Zippori were conducted by the Israeli Antiquities Authority between 2011 and 2013. The 14 excavation areas (A -N) at the site contained archaeological layers identified as Pre-Pottery Neolithic (PPNB and PPNC), Late Pottery Neolithic (Wadi Rabah, WR), and Early Bronze Age (EBA) Ib and II. Other sporadic finds were collected from top soil layers and were dated to the Roman, Byzantine and Crusader periods (Agam et al. 2016). Recycled artefacts from the two main occupations, Late Pottery Neolithic WR and EBA Ib and II were examined (see Parush et al. 2018 for details about site location and excavated areas). Most recycled artefacts came from WR layers, but artefacts from EBA contexts at the site are included in the microwear sample, which is treated as one assemblage. More detailed contextual analyses can be conducted in the future. A single blank from a PPNB layer was also examined. It is listed at the bottom of Table 1, and had no visible microwear. Artefacts in the microwear sample came from six of the excavated areas: B, C, E, K2, L, and N2 (Table 1, Figure 1). Most were from Areas B, C, and E at Ein-Zippori. Figure 1. Percentage of the 25 cores-on-flakes (left) and the 44 blanks struck from COFs (right) in the microwear sample collected from six areas of the the Ein-Zippori site. There were no COFs from Area N2. (2018)

Materials and methods
A preliminary examination of a sample of 10 recycled artefacts was conducted in one of the archaeology labs at Tel Aviv University in August, 2014. Eight of these 10 are included in the larger microwear sample of 67 recycled lithic artefacts analyzed at Ohio State University (total microwear sample size is 69 recycled artefacts). They were examined to determine: (1) if they were utilized or hafted, (2) which edges were used, and (3) what materials were worked. "High-power" incident light methods developed by Semenov (1964) and Keeley (1980) were employed. Both "low-power" and "high-power" techniques provide information, but low-power methods do not allow discrimination between specific types of worked materials (Juel Jensen 1988;Longo and Skakun 2008;Yerkes and Kardulias 1993). In Tel Aviv, the artefacts were washed in soap and water, and examined with an incident light Nikon Optiphot metallurgical microscope at magnifications of 50x, 100x, 200x (see Keeley 1980;van Gijn 1990;and Yerkes et al. 2014 for more details on the methods of microwear analysis). The 10 artefacts in the preliminary analysis came from areas B, C, E, and L at the Ein-Zippori site (eight were from WR levels, 2 were from EBA levels). Nine of the 10 had visible microwear traces, although the use wear was not well-developed. It was decided to expand the sample, so a total of 67 artefacts were shipped to Ohio State University for further examination and microphotographic documentation. One COF examined in Tel Aviv (from Area B, Locus 3538, Bucket 14858, item #2, Wadi Rabah), and a double-bulb (D.B.) lateral flake struck from a COF (from Area C, Locus 1996, Bucket 90071, also WR) were examined in Tel Aviv, but not shipped to Columbus. See the typology section (2.1) below for definitions of the types of blanks struck from COFs. In the microwear sample of 69 recycled artefacts, 56 were from WR levels, and 12 were from EBA contexts (Table 1). There also was one Kombewa type blank from a PPNB layer. At Ohio State University, the recycled artefacts were cleaned with detergent in a sonic cleaner and examined at magnifications between 62x and 375x with an Olympus BHM incident light microscope. Microphotographs were taken with a digital camera attachment at a magnification of 187.5x. Microwear traces on the Ein-Zippori artefacts were compared with use wear patterns on flaked stone tools in a reference collection of over 165 experimental tools made of a variety of different chert and flint types. The replica tools had been used for a variety of tasks including scraping, piercing, and cutting dry deer hide; scraping, sawing, and engraving bone and antler; cutting and scraping fresh hide and meat; scraping stone; and chopping, planning, whittling, sawing, and scraping wood. The functions of the Ein-Zippori artefacts were determined by matching microwear traces on their edges with use wear patterns on experimental replicas, or on published photomicrographs of wear traces on other types of experimental tools (e.g., van Gijn 1990;Keeley 1980).

Typology
A typology for flake blanks and the recycled cores-on-flakes (COFs) from which the recycled blanks were detached was established during technological analysis of lithic artefacts from late Lower Paleolithic levels in Qesem Cave, Israel. Lemorini et al. (2015) and Parush et al. (2015) used "Cores-on-Flakes/Flaked Flakes (COF-FFs)" to refer to these parent flakes. A shorter designation (COFs) for cores-on-flakes will be used in this study. Descriptions of the types of blanks made from COFs have also been modified slightly for the Ein-Zippori material (see Parush et al. 2018).
As noted above, a "parent" flake from which smaller flakes were removed is classified as a core-on-flake (COF). Subtypes of COFs are defined by the location of the detached flakes (removals). Only one of 3 COF subtypes was included in the Ein-Zippori microwear sample. All 25 were COFs with flakes or blades removed from their ventral face (ventral removals). (2018)  Of the 35 COFs with ventral removals, 20 (75%) were from WR levels, 5 (25%) were from EBA contexts. In the larger technological sample of recycled artefacts from Ein-Zippori, there were 301 COF, and 262 (87%) were from WR contexts, while 39 (13%) were from EBA levels. A total of 226 of the 301 COFs (75%) were ventral removals, while 12% (n=37) were COFs with flakes removed from the ventral and dorsal faces (combined ventral and dorsal removals), and 13% (n=38) were COFs with flakes removed from the dorsal face (dorsal removals). The last two subtypes of COFs were not included in the microwear sample. It is difficult to determine if COFs with dorsal removals are actual COFs. COFs with combined ventral and dorsal removals were also much less common than COFs with ventral removals (Parush et al. 2018). Four types of flakes or blades were removed from the ventral face or lateral edges of COFs. Blanks removed from the dorsal face of COFs are difficult to identify and were not included in the sample (Figure 2). Two types of blanks were removed from the ventral face of COFs (and have two ventral faces, one on the COF and one on the detached flake): regular flakes, and double bulb Kombewa flakes that have two bulbs of percussion (one on the COF and one on their ventral face). Blades or flakes removed from the lateral edges of COFs also have two ventral faces, and include lateral flakes that retain one of the lateral edges of the parent COF, and double-bulb lateral flakes, which have bulbs on both ventral faces and also include part of the lateral edge of the COF. A fifth type of blank called the "Tabun snap" (Shifroni & Ronen 2000) includes flakes struck from thick COFs. In the microwear sample, 36 of the 44 blanks (82%) were from WR contexts and 7 (16%) were from EBA levels (one Kombewa blank, 2%, was from a PPNB layer). In the larger technological sample of 330 flakes or blades, 84% came from WR levels and 16% were from EBA contexts. Nearly half (48%, n=20) of the blanks in the microwear sample were regular Journal of Lithic Studies (2018)  flakes, and these were also the most common type in the larger technological sample (53%, n=176), but examples of all five types were included in the microwear study (Table 1, Figure  2). However, the percentage of double-bulb lateral and Tabun snap flakes were higher in the microwear sample (10% and 14% vs. 4% and 6%) and the percentage of lateral (double ventral) flakes was lower (19% vs. 25%). See Barkai et al. (2010), Lemorini et al. (2015), and Parush et al. (2015; for more details on the typology and discussion of the technological details of recycled flake production.

Cores-on-Flakes (COFs)
As noted above, cores-on-flakes (COFs) may have microwear traces that formed during their initial use before they were discarded and later recycled. The focus of this study is on the flake blanks struck from them. However, microwear was visible on 3 cores-on-flakes (12% of 25 examined, Table 1). All three are from WR layers (15% of 20 WR COFs). None of the 5 EBA COFs had visible microwear. This is comparable to results of microwear analysis of Lower Paleolithic recycled flakes from Qesem Cave, Israel, where 15% of examined COFs were utilized (Lemorini et al. 2014). Lemorini suggested that parent flakes were not the desired product in recycled flake production. This seems to have also been the case at Ein-Zippori.
The 3 utilized COFs from Ein-Zippori included one with a lateral snap at its distal end. A lateral snap, or bending fracture, is a common type of flint knapping production failure produced by force exerted perpendicular to the dorsal and ventral faces of the core-on-flake (Johnson 1981:47;Purdy 1975:134). Lateral snaps resemble step terminations (Andrefsky 1998). The COF with the lateral snap termination was from Area B, Locus 3538, Bucket 14858 (item#1). This steep edge (edge angle, E.A. = 60 ᵒ ) was probably used to scrape meat or fresh hide off of bone (light butchering). There was very little use wear, but bone seems to have been nicked with the tool producing some small patches of bone microwear polish. One of the lateral edges of this COF was also used for cutting meat or fresh hide ( Figure 3A, a, b, c). The micowear was not well developed, and there was very little edge damage. Another COF with ventral flake removals had a small patch of dry hide and stone-on-stone microwear on one of its steep lateral edges (E.A. = 70 ᵒ ). It seems to have been used as an expedient hide and stone scraping tool ( Figure 3B, d). The third utilized COF was examined in Tel Aviv (not illustrated). It had a few meat or fresh hide and bone traces along one of its edges. The orientation of the microwear suggested cutting and scraping. Possible prehension traces were also recorded (Table 1). Like the COF with the lateral snap (above), this COF from the same Area (B), Locus (3538), and Bucket (14858, it is item #2) seems to have been used as an expedient butchering tool. (2018)

Blanks struck from core-on-flakes (COFs)
Of the 44 blanks removed from COFs in the Ein-Zippori entire microwear sample, 19, or 43% were utilized (Table 1, Figure 4). Only 2 of the 19 (10%) were from EBA levels, one of these was used for light butchering and the other to cut, scrape, and whittle bone or wood. Of the 17 utilized WR blanks (90% of the total), 7 (41%) seem to have been used for light butchering (cutting meat and fresh hide and scraping meat off of bones). Four (24%) were used as wood-working scrapers, gravers, and knives. Three utilized WR blanks were used to cut, scrape, and whittle bone or wood (18%). Two were used to cut and scrape meat and soft wood (12%). One blank, from the same Area (E), Locus (3135), and Bucket (30816), as a flake used for light butchering, only had weakly-developed "generic" microwear traces (5%) that may have formed when it was used to cut or scrape meat, but it was not certain that it was used for butchering (Table 1, Figure 4). . Four (21%) were used for wood-working, 4 others (21%) were used to cut, scrape, and whittle bone or wood, and 2 (11%) were used to cut or scrape meat and soft wood. One had "generic" microwear traces that could not be associated with a specific task.
All 19 utilized blanks had edges with rather sharp edge-angles (E.A.) of 20 ᵒ -45 ᵒ , but most of the sharper edges were used for cutting, while the steeper edges were used for scraping or whittling. None of the utilized blanks had visible hafting traces (Figure 4). The single flake that could not be classified (listed as varia on Table 1) was not utilized.
Microwear traces were visble on all five types of blanks struck from COFs (Table 1). Each type of blank was used for several different tasks. Regular flakes were the most common form, and 9 of the 20 regular flakes in the Ein-Zippori sample (45%) were utilized. Eight were from WR levels, one was from an EBA context (Table 1). Not all of the utilized Regular flakes are illustrated, but 5 (Figures 5 B, C & D; 6 A) were used for light butchering (56%), 2 were used to cut meat and work soft wood (22%, Figure 7A & B, Table 1). Two others, one from an EBA (Figure 8 A) and one from a WR level (Figure 8 B), were used to cut and scrape (or whittle) bone or wood (22%). (2018)    Wear traces were also identified on 2 of the 4 Kombewa type flakes (50%). One from an EBA level was used for light butchering (Figure 6 B). Microwear on the other (WR) blank was not well-developed. Apparently it was used for scraping or whittling soft wood (Figure 7  D). Four of 8 lateral flakes were utilized (50%), all from WR layers. One was used for light butchering (Figure 6 D). Two were used to cut, scrape and engrave wood (Figure 8 E). The fourth showed only generic weak microwear traces (possibly meat). Two of 4 double-bulb (D.B.) lateral flakes also had microwear (50%). Both were used to cut, scrape, and whittle bone or wood (Figure 8 C&D).

Journal of Lithic Studies
Of the six WR Tabun snap flakes in the samples, two were utilized (33%). One was used for cutting and scraping meat or fresh hide and bone (or light butchering). This use wear was visible on a utilized edge above the Tabun snap (Figure 6 C & c). The use wear on the other was not well-developed. Apparently it was used for scraping or whittling soft wood (Figure 7 C).
Wear that may have been cause by prehension (holding the flakes tightly when they were being used) was recorded on 6 flakes: 3 regular flakes, 2 double-bulb lateral, and one Tabun snap. It is likely that all of the utilized flakes were held in the hand when used, since no clear hafting traces were noted on any of them (see Rots  One Kombewa flake with "natural luster" on its edges is also shown ( Figure 5 A & a). Many of the COFs and blanks in the Ein-Zippori sample had lustrous surfaces, and at times it was difficult to recognize the microwear traces. However, even though the utilized recycled COFs and blanks had only been used for a short time, distinctive polish and striation patterns similar to those on experimental stone tools were visible, and it was possible to interprete their functions. For example, Figure 5, B, C, and D are small regular flakes with sharp edges that have rough hide or meat microwear polish on their edges (Figure 5 b), or hide or meat and brighter bone polish ( Figure 5 c & d). These 3 small sharp-edged regular flakes were probably held in the hand and used as expedient butchering tools.
On Figure 6 there are four different types of blanks that were all used as expedient butchering tools, regular (Figure 6 A), Kombewa (6 B), Tabun snap (6 C), and lateral (6 D). The variation in these blank types that were all used for the same task illustrates the expedient use of small recycled flakes. Any flake with thin razor-sharp edges struck from a parent COF was selected for meat and hide cutting and for scraping meat off of bones. No specific type of blank was produced only to be used for the light butchering.
On Figure 8, two regular flakes (A and B) that were used to cut, scrape, or whittle bone, wood, or both are illustrated. Some of the edge-angles (30 ᵒ and 40 ᵒ ) on these regular flake blanks are steeper than the razor-edged blanks shown in Figure 6. They seem to have been more versatile flake tools used for bone or wood scraping and whittling. A wedge-shaped double-bulb lateral flake blank (Figure 8 C) was also used for bone or wood scraping and whittling. The distal end of this blank could have been used as a graver to cut grooves in wood, bone, or antler objects, but there were no visible wear traces on that part of the blank. However, bone or wood microwear (Figure 8 g & h) was visible on graver edges of another double-bulb lateral flake (Figure 8 D), and on a lateral flake (Figure 8 E). Possible prehension microwear on one lateral edge of the double-bulb lateral flake (7 D) suggested that it had been held in the hand when it was used. The lateral flake (7 E) was worn, but it also seems to have been used to scrape soft wood or bone (Figure 8 i). Possible prehension traces were also visible on this flake.
On Figure 7, two regular flakes A and B are shown with weakly-developed microwear traces that suggest that they may have been used to cut and scrape both meat and soft wood (Figure 7 (Figure 7 D) were near its "graver-like" distal end. While the microwear traces are not well-developed, these two blanks with pointed ends (Figure 7 C & D) look like they could have been used as gravers, but instead were used to scrape or whittle wood.

Conclusions
Microwear traces on many of the 19 utilized blanks struck from recycled COFs found at Ein-Zippori were quite similar to the wear traces that Lemorini et al. (2015) observed on recycled Paleolithic artefacts from Qesem Cave. The micowear on the blanks struck from COFs is from new tasks that were performed with them. In the microwear sample of 134 recycled blanks from the late Lower Paleolithic contexts at Qesem Cave, most regular flake blanks had been used for cutting meat while many lateral flakes were used to cut plants (Lemorini et al. 2015). The small thin regular flakes were also the most common type in the microwear sample from Ein-Zippori, and while 56% of them were used for light butchering, they were also used for bone-working (22%) and wood-working (22%). The other four blank types that were utilized, lateral flakes, Kombewa flakes, double-bulb lateral flakes, and Tabun snap flakes, also were used for light butchering, meat cutting, and bone and wood working. At Ein-Zippori, there does not seem to be a strong correspondence between the form of the blanks struck from COFs and their functions.
"Older" flakes from earlier activities and occupation episodes were recycled during the Paleolithic at Qesem Cave, and during the Late Pottery Neolithic and Early Bronze Age periods at Ein-Zippori. The products of recycling were expedient flake tools that were made and used in an ad hoc fashion. However, the way that the blanks were used by the WR and EBA settlers at Ein-Zippori was a little different. The majority of the recycled blanks examined in this study came from WR levels (82%), and 47% of the WR blanks were utilized, while only 28% of the EBA blanks in the sample had been used, so the picture of recycling during the Wadi Rabah phases is clearer. Nonetheless, both WR and EBA inhabitants of Ein-Zippori did not produce specific types of small blanks from recycled COFs that they used for specific tasks like mobile Paleolithic foragers had done at a much earlier time. They made and used several new types stone tools in their farming, herding, and craft activities. Their use of recycled flakes seems to have been for light butchering and wood and bone working when they did not have their flaked stone or metal knives and scrapers with them. They used the discarded flakes they found lying around as cores and produced sharp-edged flakes to cut meat or to scrape or whittle bone or wood. These activities seem to have occurred at many locations across the settlement and do not seem to be associated with specific butchering or crafting areas.