MICROARTIFACT ANALYSIS AT TEL HALIF, 1992 SEASON Arlene Miller Rosen Archaeological Division Ben Gurion University Beer Sheva, Israel The objective of this research was to analyze the microarti­facts embedded in room-floors in order to identify evidence of activities performed in the rooms by their ancient inhabitants. Microartifacts (defined here as archaeological remains from 30-0.25 mm in size), are not usually removed from room floors, and are resistant to sweeping because of their small size. They accumulate on a surface during the life of a room, and form a record of activities within that living area. METHODS Samples were taken from all room floors excavated in the 1992 field season from both Field IV, and Site 101. A detailed description of the methods can be found in Rosen 1989, 1991, but a brief account is as follows. In the case of packed mud floors, every effort was made to select a section of the surface that was well-preserved. In many cases, the best sections of flooring were found under large broken sherds. Other areas were also sampled to avoid the bias of always sampling under large sherds which were usually in the vicinity of storage vessels. On cobbled surfaces, two types of samples were collected. One type was from the ashy deposits immediately overlying the cobbles, and the other was the sediment in which the cobbles were embedded. We did this in order to locate the sediments best representative of activities on the surface. The ashy deposits above the cobbles were in some cases (e.g. Areas F8 and G8), in situ remains of activities, distinct from other ashy remains of destruction collapse. This was determined by the lack of wood charcoal, abundance of seed charcoal, and the remains of other foodstuff such as fish bone and bird eggshell. The samples were taken by scraping a ca. 25 cm surface area down to a depth of ca. 1 cm and collecting the resulting sediment in a clean plastic bag. The material was washed through a 0.25 mm sieve, allowed to dry, and then resieved through a set of nested sieves resulting in six size fractions of 30-10 mm, >5.00 mm, >2.00 mm, >1.00 mm, >0.5 mm and >0.25 mm. Each size fraction was examined under a stereo-binocular microscope, and the per­centage of different categories of artifacts was determined with the help of visual percentage charts. The resulting data were divided into two categories represented by Table 1 and Table 2. Table 1 displays the data as percentages of microartifact catego­ries by size fraction for each sample, and Table 2 shows total percentages of microartifact types by sample. PRELIMINARY RESULTS Samples were taken from all temporal phases of the excava­tion, and the results appear in Table 1 and 2. However, since we are concerned here with deposits from living floors, and there were no clear occupation surfaces in the Persian phase, this discussion will focus primarily on results of analyses from the Iron II levels of Field IV, and the Chalcolithic surfaces of Site 101. Field IV For the purposes of this study, the room floors excavated in different squares will be numbered consecutively from N to S. Graphs of total percentages of the most common micro-artifacts, Charcoal, Bone, Flint, and Sherds are plotted on room plans in Figures 1--5. Other significant micro--artifacts are also noted on these plans. Room 1 (Area F8, Fig. 1): Only part of the room appeared in the excavation square. Two samples were collected from the pale gray ash directly overlying a cobbled surface. These contained the remains of kitchen activities. Sample 33 included fish bone and a piece of a beachrock mortar. Just south of this, in Sample 32, there were the remains of cereals, a flint chip with sheen, (perhaps a sharpening chip from a sickle), and the burnt bones of a small rodent that may have been associated with the structure at the time of destruction. The southernmost sample, 32, has the highest concentration of food remains (charcoal and bone). Room 2 (F8/G8 baulk and Area G8, Fig. 1): This room also consisted of a cobbled surface, and was almost complete. Seven samples were collected here, also from the ashy deposits overly­ing the cobbles. Again, this ash was debris from occupation, with very little attributable to destruction collapse. Most of the charcoal was from charred seeds, either whole or broken, and very little was from wood. This suggests that food preparation was conducted in the room, but cooing was done elsewhere. This is supported by the lack of a hearth or _Tabun_ in the room. The presence of a"cultic corner"in the room may indicate that some of the food remains served as offerings. The most abundant food remains in this room come from within the triangular area formed by the small altar, the large flat stone, and the incense burner. The seeds from this sample (28) are dominated by grape pips, with some cereals and legumes. Fish bones come from samples 1, 27, and 28. Sample 10 contained a fragment of a beachrock mortar, further indication of food preparation. Room 3 (Area H8, Fig. 2): The mud surface of this room was only preserved in patches in the southeastern quadrant of the square. Five samples were collected from this area. Although the absolute percentages of charcoal were low, the charcoal was primarily from seeds (cereals and legumes), except in samples 15 and 16. Sample 15 had a relatively high percentage of bone from medium or large mammals (i.e., food remains, as opposed to small intrusive mammal bones). It also contained bits of bitumen, perhaps from the haft of a stone tool. The adjacent Sample 14 contained a small retouched flint flake, or "expedient" tool and some pieces of debitage. Samples 13, 15, 16, and 25 all con­tained remains of fish. Samples 13-16 had a high percentage of sherds in the small size fractions which is an indicator of heavy foot traffic in this room, either from a long period of occupa­tion by a few people, or broad access to the room by a number of people. Room 4A and 4B (Area I8, Fig. 3): This includes the bin containing loom weights (4A) and the adjacent cobbled surface to the south (4B). Two samples were collected here, one at each subdivision of the room. Sample 30 from the bin was collected at a point too high above the occupation surface and probably repre­sents brick collapse. It has very few microartifacts, and the very small amount of charcoal is all from wood. Sample 37 from the 4B cobble surface is also poor in microartifacts but does contain seed charcoal and fish bone. Room 5 (Area J8, Fig. 3): The boundaries of this room are undefined archaeologically, but the packed mud floor is clear, and was marked by a large quantity of flat-lying sherds. The nature of the floor sediment itself is characteristic of prepared surfaces. It is clayey in texture and grayish in color with tiny flecks of charcoal embedded within. An examination of the sediments at the base of the fenestrated basalt bowl that was embed­ded within this surface showed that there are other floor levels beneath the uppermost one (J8010). Two samples (40, 47) were collected from this surface. Sample 47 is as yet unanalyzed. Sample 40 contains a high percentage of charcoal from both wood and seeds, as well as the inevitable fish bone. The wood char­coal could either originate from an as yet unidentified hearth area, or it could signify mixture with the remains of roofing material destroyed in the destruction. It would be useful, in the following season, to excavate the floor levels just underly­ing J8010 in order to compare material from earlier phases of the same structure. Room 6A and 6B (Area K8 and K8/K9 baulk): These represent the cobbled surfaces in the northern section of the pillared house. One sample (43 was collected from Room 6A, the parti­tioned store room in the NW quadrant of the house, and one sample (41) comes from Room 6B, the main portion of this cobbled sur­face. Both samples have numerous phytoliths (in the less than 0.25 mm fraction) and impressions of stems and plant floral parts in clay. A quick examination of the phytoliths in the field showed that the plants were from wild grasses and weeds. The stem impressions were not aligned as would be expected in mat or basket remains. Therefore, this concentration of wild plants is probably related to a floor preparation technique which consisted of a thick layer of plant matter spread onto the surface and "glued" together with clay. Both samples were poor in microarti­facts, but Sample 43 did contain some seed charcoal, and both had fish bone. Room 7A and 7B (Areas K8, K8/L8 baulk, L8): This is the room (7B) and storage area (7A) just south of the pillars, that con­tain a mud floor and large flint saddle quern. The richest sample from this room is Sample 24 which was associated with what may be a collapsed Tabun. Pieces of a restorable cooking pot were also in this location. The sample contained Tabun fragments and a large percentage of charcoal consisting primarily of cereal grains, with some legumes. There were also bird eggshell and fish bone. A likely reconstruction of this sample is that most of the remains came from within the Tabun and are the remnants of cooked meals. The dominant cereal grains could represent the contents of the cooking pot (some sort of gruel?) left at the time of abandonment. Sample 45, just South of the flint quern has a fragment of beachrock mortar, suggesting that other types of mortars were used in this area as well, either predating or contemporary with the flint quern. This sample also contained a fragment of bone with a very high sheen, similar to the weaving shuttles recovered in other locations at the site. Sample 46 comes from a small storage area and is distinguished by the same type of surface preparation technique (weed stems and floral parts impressed in clay) as was found in Rooms 6A and 6B. There are also fish bones and seed charcoal. Room 8 (Area L8): This is the cobbled surface just south of Room 7. The one sample analyzed from this surface (36) is domi­nated by wood charcoal and may represent destruction collapse, although one fish bone was also recovered. Some additional isolated samples include two samples from Area G7 (Fig.5). Sample 49 was taken from the NW quadrant of the square and was associated with area around the pillar base and krater. It contained primarily food remains, including carbon­ized cereals, bird eggshell, and, of course, fish bone. The other sample (42) was collected from the baulk in the SE quadrant of the square, and represents material overlying the cobbled surface. It did not contain a large quantity of microartifacts, but did have a small proportion of cereal grains and fish bone. The sample from the living surface in F7, is one of the two as yet unanalyzed samples. Site 101 Samples taken from Site 101 (Figure 6) generally have higher percentages of bone and flint than those from Field IV. There is also a wider range of microartifact types within a restricted area. These include dense concentrations of phytoliths from the glumes or seed coating of wheat, bitumen, beachrock mortar frag­ments, copper minerals, fish bone, red ocher, mica, and bird eggshell. This suggests a very unspecialized use of the living floors, with each area of the room available for any household or craft activity. Discussion The results from the 1992 season at Tel Halif are still preliminary, since the majority of the rooms excavated were not yet in a clear architectural context. Further work in coming seasons should delineate the character of the buildings associat­ed with these floors. However, it is still possible to make some general comments. In Field IV, the microartifacts (as well as the conventional artifacts) reflect activities that are primarily associated with the storing, processing, and cooking of various foods. These food-related tasks are presumably female activities, and it is likely that most of the excavated floors in the Iron II phase this season are related to female domestic work areas. It is interesting, then, to consider the cultic remains (altars, incense burner and pillar figurine) from Area G8/F8, including the concentration of charred grape pips associated with them as possibly related to a female cultic practice. In Rooms 6-7, the rooms are partitioned into discrete areas which give the impression of specialized uses for each portion of the house. This contrasts markedly with the picture from the Chalcolithic surfaces which contain remains from a wide variety of activities in a relatively small area. These activities prob­ably represent both female and male work areas together. With expanded excavations in future seasons we hope to broaden the contexts from which microartifacts are analyzed to include other types of activity and work areas. Microartifact Analysis from the 1993 Season at Tel Halif Arlene M. Rosen Archaeological Division Ben Gurion University Beer Sheva, Israel The goals for microartifact analysis for the 1993 season at Tel Halif were similar to those from the previous 1992 season. We were interested in using microartifacts to understand the activities conducted on archaeological surfaces and the functions of rooms within the excavated structures. The three periods represented in the samples from this season were Persian, Iron Age, and Chalcolithic. The two Persian period samples were collected from a possible floor, and the base of a bin in Field IV, the Iron Age samples came from the intact floors of domestic structures in Field IV, and the Chalcolithic material was taken from the lowermost living surfaces of the cave dwelling in site 101. Results of the analyses are listed in Table 1, representing the microartifacts broken down into percentages by fraction size for each sample, and Table 2, the total percentages of microartifact categories per sample. The most significant results are outlined below by Field and period. FIELD IV Due to the nature of preservation in this type of site, some excavation squares opened up a broad expanse of archaeological floors and surfaces, others were more patchy. In the broad exposures of room floors it was possible to take a number of samples across the surface, and the results show patterning in the concentrations of different types of microartifacts. On the smaller surfaces where only one sample was collected, we can only give a description of the microartifacts from the one point. One of the best preserved suite of floor samples comes from Area F7 (including the baulks of F7/F8 and F7/G7) which exposed a large area of surface F7006 and parts of surfaces F7018, all from stratum VIB1A (see Figure 1). Only one sample, H-93-22, came from floor F7018 due to the limited size of exposure. The most notable feature of this sample was a relatively high percentage of sandstone mortar fragments, from two different types of sandstone. This suggests that there were two generations of mortars used in the room, and a long­term use of this area for grinding. The presence of seed charcoal also suggests the preparation of grains. The percentage of sherd fragments in the micro-fractions can be taken as a measure of the amount of trampling and traffic within a given room. When compared to other rooms, the sherd percentage in this sample is relatively low, perhaps indicating somewhat limited access to the room. The bone fraction includes fragments of large bones and fish, which are also the remnants of food items. The sample to the west of wall F7016 (H-93-23) contained only small amounts of microartifacts, but it did contain some olive seeds, and fish scales, and more significantly, the burned remains of rodent bones. Although rodent bones occur in many samples, the burned ones are most likely contemporary with the Iron Age occupation surface. The presence of rodents in this room might indicate a location for food storage. Surface F7006 is notable for the high percentages of sherds in the fine fractions which indicates a great deal of trampling traffic in this room suggestive of free access and intensive use. This room clearly had a domestic function and a variety of different uses. Charcoal from all of the samples contained both wood and seeds including cereals and legumes indicating food preparation and consumption. The bones come from fragments of large bones (from meat consumption) as well as a large percentage of fish bones and scales. The correlation coefficient for charcoal and bone from the F7 sample is high, (0.86) (see Table 3 and Figure 2) which is characteristic of areas of food preparation (see Rosen 1989, 1991). Sample H-93-12, is notable for a quantity of homogeneous brown flint chips from the production or sharpening of a tool at that location. Samples H-93-24,-25,­27, and -28 which cluster in the southwest portion of Area F7, all contain significant percentages of greenish slag which could indicate the proximity to some sort of domestic kiln. Surface F8009 is in the room adjoining that of surface F7006. It is represented by Samples H-93-37, -38, and -40, which are notable for their high percentages of bones, cereal remains and flint chips. Again the correlation coefficient for bones and charcoal is very high (0.99) indicating a food preparation function. Another well-preserved surface, H7005 from Area H7, has a very different character from both F8009 and F7006 (see Figure 3). Although there is a small amount of seed charcoal and some fish bones, the overall character of the surface is not characteristic of food processing and consumption due to the lack of correlation between the charcoal and bone remains (see Table 3 and Figure 4). Three of the samples (H-93-13, H-93-14, and H-93-17) are notable for their high percentages of greenish-colored slag, and the presence of some red ochre. The correlation coefficient for all the microartifacts on this surface was highest for the sherd/slag relationship (O.85). This suggests that these samples are in the vicinity of an as yet uncovered, or unpreserved kiln, perhaps for the manufacture of pottery. The ochre is unusual in the Field IV samples and may represent pigment for pottery. The other samples from Field IV generally come from floors which were only partially preserved. Due to the small exposures, just one sample was taken per surface. Therefore, it is only possible to generally describe their contents without a discussion of patterns. Sample H-93-19, from Area I7 (Locus I7006), contains some seed charcoal, fish bone, and flint chips, but also includes a high percentage of plaster which is unique. It indicates that either the floor itself, or the walls of the room were plastered, suggesting a non-domestic (perhaps elite) function for the room. In Area J8, one sample (H-93-39) was taken from J8010, a continuation of a surface excavated in the 1992 season. Unlike the samples from F7006 and F8009, the high percentage of sherds is probably more indicative of the broken vessels found on the surface then of trampling, since most of the sherds fall into the larger size fraction (>10 mm) (see Table 2). The sample includes fish bone and some type of bird eggshell which is thicker than modern chicken eggshells. This might be the remains of food goods. Sample H-93-29 comes from the fragment of floor (K9001) in K9, close to the iron implement recovered from that locus. It contains a modest amount of cereal and bone fragments, similar to the other samples but also reflects the presence of the iron implement by a relative high percentage of iron fragments ("other" category). This is significant from a methodological perspective since it suggests that the microarchaeological technique could have detected the presence of an iron implement even in its absence. Sample H-93-31 comes from sediment overlying cobbles in the west room of the pillared building (Locus L9006). The results suggest that this surface is either not a true surface, or it is contaminated with collapse debris. This is indicated by the grain-size distribution of charcoal (decreasing percentages with decreasing size, a pattern indicative of fills), and the minute quantity of bone (generally ubiquitous on true floors). Finally, Sample H-93-32 comes from Locus G6014 at the base of the Persian (?) period silo. It contains a low percentage of sherds, a small percentage of cereal grains, bitumen, eggshell, fish scales, and fish vertebrae. All of these items could be the remains of midden material tossed into the pit after it had gone out of use, and do not seem to reflect the function of the bin. SITE 101 A total of 16 samples were processed from site 101. As in the samples from the 1992 season, the diversity of microartifacts is much richer in the Chalcolithic levels than in the Iron Age from Field IV. This is interpreted as a function of the greater diversity of daily tasks in each Chalcolithic household, with respect to the Iron Age households. Locus 91081 Eleven of the samples come from surface 91081 (see Figure 5). They are notably low in charcoal, with a few exceptions including isolated seeds from cereals or legumes. The lack of correlation between the charcoal and bone groups suggests that the cooking and consumption of food was done elsewhere, perhaps outside of the cave. Although the presence of phytoliths from these surfaces does indicate that grain was probably stored close to this location. Bone percentages are also generally low with the marked exception of Samples H-93­-18 and H-93-35. The bone is composed of burned and unburned fragments of large bones. A very small amount of fish bone occurs in H-93-3. There is a very high correlation between bone and flint chips in these samples which is especially evident in sample H-93-18 (see Table 3). This implies that the activity in the vicinity of sample H-93-18 involved both bone and flint, either the cutting of meat and bone, or perhaps the manufacture of bone tools. Sherd percentages are highest along the line formed by Samples H-93-1, -5, and -30. Both Samples H-93-15 and H-93-36 have sizable fragments of mortars, mother-of-pearl, and ochre. The occurrence of these three items at two different locations suggest the similarity of some activity, perhaps the grinding and production of pigments. Bitumen occurs only in Sample H~93-4 from this surface. Metallic slag and copper ore is present in both Samples H-93-5 and -30, and appear to be a waste deposit from a metal working operation. Sample H-93-16 contains a high percentage of lime plaster flakes with straw inclusions. Perhaps these are the remains of a plastered area of floor. Only three samples come from Surface 91076 (H-93-6, -7, and -8). They also have low percentage of bone and charcoal, and the most significant feature is the bitumen and sandstone mortar fragment in H-93-7, and the fragment of spiral sea shell in H-93-8. Similar to Surface 91081, these samples show a high correlation between bone and flint chips. The last two samples are from Surfaces 90041 and 101051. Both are similarly poor in microartifacts, but both contain a greenish slag, different from the metallic slags of Surface 91081. References Rosen, A.M. (1989). Ancient town and city sites: A view through the microscope. American Antiquity 54(3):564-578. Rosen, A.M. (1991). Microartifacts and the study of ancient societies. Biblical Archaeoloqist 54:97-103.