Tool Blank Selection, Debitage and Cores from Abu Hureyra 1, Northern Syria

PaléorientCNRS Editions

TOOL BLANK SELECTION, DEBITAGE AND CORES FROM ABU HUREYRA 1, NORTHERN SYRIAAuthor(s): Deborah I. OLSZEWSKISource: Paléorient, Vol. 15, No. 2 (1989), pp. 29-37Published by: Paléorient and CNRS EditionsStable URL: http://www.jstor.org/stable/41492360 .

Accessed: 27/01/2015 12:36

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Paléorient and CNRS Editions are collaborating with JSTOR to digitize, preserve and extend access toPaléorient.

http://www.jstor.org

This content downloaded from 130.91.82.177 on Tue, 27 Jan 2015 12:36:07 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=paleorient

http://www.jstor.org/action/showPublisher?publisherCode=cnrseditions

http://www.jstor.org/stable/41492360?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


PALÉORIENT, vol. 15/2 - 1989

TOOL BLANK SELECTION, DEBITAGE AND CORES

FROM ABU HUREYRA 1, NORTHERN SYRIA

D.I. OLSZEWSKI

ABSTRACT. - An analysis of tool blank selection, debitage and cores from the Epipaleolithic levels at Abu Hureyra shows that several chronological changes occur. The most important of these is a reduction in the production of flakes, and consequently less use of flakes as tool blanks through time. The latest levels at Abu Hureyra 1, those of Unit 4, appear to be roughly contemporary with Phase la from Tell Mureybet, and point to similarities, as well as differences, in the occupations at these two sites at the end of the Epipaleolithic period in the Levant. RÉSUMÉ. - L'analyse de la sélection des supports d'outils, des produits de débitage ainsi que des nucléus en provenance des niveaux épipaléolithiques d'Abu Hureyra indique plusieurs changements chronologiques. Le plus important d'entre eux est une baisse de la production d'éclats et donc une utilisation réduite des éclats comme supports d'outils au fil du temps. Les niveaux les plus récents d'Abu Hureyra 1, ceux de l'unité 4, semblent à peu près contemporains de la Phase la de Tell Mureybet , et mettent en évidence aussi bien des similitudes que des différences au cours des occupations successives de ces deux sites à la fin de l'Epipaléolithique au Levant.

INTRODUCTION

Abu Hureyra 1 represents the Epipaleolithic occupation at Tell Abu Hureyra, which is located along the Middle Euphrates River in northern Syria (fig. 1). This occupation consists of deposits located in the lower part of Trench E. An unknown amount of the upper portions of the Epipaleolithic occupations were removed by leveling during the construction of Neolithic buildings in the subsequent settlement of Abu Hureyra 2. The Abu Hu- reyra 2 occupation during the Neolithic period occurs after an apparent abandonment of this area of the site for approximately 500 years (1).

The series of settlements representing Abu Hu- reyra 1 span approximately a millennium and a half, from 11,500 to 10,000 B.P. (2). These dates indicate that the occupations are probably partially contemporary with the Early Natufían of the southern Levant, from 12,500 to 1 1,000 B.P. (3), and certainly overlap with the Late and Final Natufían of that region, from 11,000 to 10,500 B.P. or later (4). Most important, however, is the fact that the dates from Abu Hureyra 1 indicate that the later Epipaleolithic settlements here are partially contemporary with Phase la ("Final Natufían") and possibly also with Phase lb ("Epi-Natufian") at Tell Mureybet, the only other extensively excavated representative of this period along the Euphrates River (5).

Recent opinion has been divided regarding affini- ties between the north Syrian Late Epipaleolithic chipped stone assemblages and the Levantine Na- tufían (6). This paper, however, will concentrate on an examination of the cores, debitage and tool blanks found at Abu Hureyra 1, and a comparison of these to the data from Tell Mureybet. This allows the Middle Euphrates River area of northern Syria to be assessed as a specific region, and provides the only detailed comparison of these two late Epipaleo- lithic sequences.

(1) MOORE, 1975; 1979; MOORE et al., in preparation. (2) MOORE et al., 1986; MOORE et al., in preparation. These

dates are only approximate, since the standard deviations must also be taken into account.

(3) BAR-YOSEF and VALLA, 1979 : 148; BAR-YOSEF, 1981 : 405; 1983 : 13; HENRY and SERVELLO, 1974 : 35-36.

(4) Cf. footnote 3. (5) CAUVIN J., 1972; 1974; 1977.

(6) BAR-YOSEF, 1981 : 399; BYRD, 1987 : 310-313; CAL- LEY, 1984; 1986 : 193-199; CAUVIN M.-C., 1980; OLSZEWSKI, 1984; 1986 : 104-108; 1988.

FIG. 1. - Location of Middle Euphrates sites discussed

29



The late Epipaleolithic occupations at Tell Mureybet and Abu Hureyra 1 document the period from about 11,500 to 10,300/10,000 B.P. in a region other than the southern Levant (Natufian complex). Units 1-3 from Abu Hureyra 1 comprise the earliest portion of the sequence. Unit 4 from Abu Hureyra 1 and Phase la from Tell Mureybet provide information on occupations dating to the end of the late Epipaleolithic, as well as interesting contrasts in site functions during a roughly contemporary period. Phase lb from Tell Mureybet incorporates information covering the transition to later Neolithic settlements.

DISCUSSION OF ABU HUREYRA 1 DATA

The late Epipaleolithic levels of Abu Hureyra 1 (Trench E) comprise a seven by seven meter excavation, the depth of which varies from about 50 cm to 1 .3 m, depending on where the underlying subsoil was encountered (7). These levels have been organi- zed, on the basis of typological frequencies of chipped stone tools, into four chronological units. The earliest of these is Uniti, the latest, Unit 4. Based on a series of uncalibrated radiocarbon dates (8), and the levels that these represent, the units can be tentatively dated as follows : Unit 1, about 11,500 to 1 1,000 B.P., Unit 2, from perhaps 11,000 to 10,600 B.P., Unit 3, from ca. 10,600 to 10,400 B.P., and finally, Unit 4, from 10,400 to perhaps as late as 10,000 B.P.

The analysis of the cores and the debitage was not focused on metric studies or specific technological attributes, such as platform types, edge configura- tions and the like, and therefore provides less technological detail than, for example, the excellent recent study for Tell Mureybet by Calley (9). Howe- ver, the Abu Hureyra 1 data are amenable to a series of technological observations that allow intrasite and intersite comparisons.

Debitage classes of flakes, blades and bladelets are divided into several types which provide a certain measure of the technological processes that occurred here. Primary decortication flakes (ldc) are completely cortical; secondary decortication flakes (2dc) exhibit less than 100 % cortex, and plain flakes are those with no cortex. Microflakes are defined as having a maximum dimension of 2.5 cm. Rejuvena- tion flakes refer to flakes that were produced when removing prior core platforms. This latter category includes specialized flakes such as core tablets

(extremely rare in the Abu Hureyra 1 collections), and a variety of less well-formed core platform removals.

Several types also comprise the blade and bladelet classes. Secondary decortication (2dc), as for flakes, refers to partially cortical blades/bladelets (by definition, primary decortication blades/bladelets do not occur). Additionally, noncortical blades/bladelets are divided into first order (1st o) with triangular cross-sections, and second order (2nd o) with quadrilateral cross-sections. Finally, rejuvenation blade/bladelets are those that mark the removal of prior core platforms. Like rejuvenation flakes, rejuvenation blades and bladelets are not well-formed, and the Abu Hureyra 1 units contain few examples of crested blade/bladelets.

TABLE la Debitage Classes from Abu Hureyra 1

CLASS/TYPE UNIT 1 UNIT 2 UNIT 3 UNIT 4 NX NX NX NX

Flakes 11,171 87.7 15,998 90.4 14,825 87.2 18,642 86.9

Blades 690 6.9 918 5.2 1,238 7.3 1,825 8.5

Bladelets 663 5.4 762 4.4 940 5.5 978 4.6

TOTAL 12.734 17.6-8 17,003 21,445

TABLE lb Debitage Types within Classes from Abu Hureyra /

CLASS/TYPE UNIT 1 UNIT 2 UNIT 3 UNIT 4 NX NX NX NX

Flakes 11,171 15,998 14,825 18,642

ldc 1,270 11.4 1,734 10.8 531 3.6 485 2.6 2dc 2.473 22.1 3,536 22.1 4,053 27.3 3,930 21.1 plain 3,355 30.0 5,773 36.1 4,670 31.5 5,125 27.5 micro 3,721 33.3 4,470 27.9 5,336 35.9 8,886 *7.7 reouv 352 3.1 485 3.0 235 1.6 216 1.1

Blades 880 918 1,238 1,825

2dc 2*6 33.6 293 31.9 302 24.4 379 20.8 1st о 181 20.6 202 22.0 314 25.4 508 27.8 2nd о 307 34.9 350 38.1 568 45.9 894 48.9 rejuv 96 10.9 73 7.9 54 4.3 44 2.4

Bladelets t>83 762 940 978

2dc 115 16.8 105 13.4 132 14.0 93 9.5 1st о 348 50.9 407 52.0 439 *6.7 497 50.0 2nd о 198 29.0 241 30.8 344 36.6 374 38.2 rejuv 22 3.2 29 3.7 25 2.6 14 1.4

TOTAL 12,734 17,b98 17,003 21,4*5

(7) MOORE, 1975; 1979; MOORE et al, in preparation; also field plans and profiles of the excavation of Trench E.

(8) MOORE et al., in preparation. Dates are preliminary approximations.

(9) CALLEY, 1984; 1986.

30



Table la provides the raw counts and percentages for the major debitage classes of flakes, blades and bladelets, as well as for types within these classes. If one examines the percentage data for the major classes, there is little apparent change through time in the debitage from the Abu Hureyra 1 units. This might suggest that the core reduction and tool blank selection strategies were similar throughout the Epipaleolithic occupational history at this site.

However, a closer examination of the types of debitage within each of the major classes (Table lb) suggests that certain changes did occur. Within the class of flakes, both primary decortication and rejuvenation flakes decrease in frequency through time, and this same trend is seen in the frequencies of secondary decortication and rejuvenation types for the blade and bladelet classes. Since these types reflect initial core reduction and core platform preparation, the temporal trend toward lower frequencies suggests a shift in chipped stone production strategies. This could simply indicate a shift in the location of initial core reduction, in the later units, to areas of the site that are unexcavated.

However, if the excavated sample is representative of the debitage over the site as a whole, the decrease in the frequency of core preparation ele- ments suggests several possible explanations. The source of raw material may have been located farther away from the site later in time, or raw material may have become less readily available through time. Alternatively, activities at the site may have changed through time.

Previous' analyses (10) have demonstrated that changes in the frequencies of certain tool classes do occur through time at Abu Hureyra 1. If these monitor changes in activities, it is possible that such changes are also reflected in the debitage types.

Some support for the notion of less readily available or more intensively used raw materials can be found in the data. Table 1 shows a large increase in microflakes by Unit 4. Since microflakes are often the product of forming or resharpening tools, an increase in this category suggests that pieces are more intensively retouched and therefore perhaps more intensively utilized in the latest unit.

Table 2 presents information on the typps of cores used to manufacture flake blanks. Generally spea- king, the most common core type is the multiple platform variety, followed by the single platform type (11). Simple single platform cores are consistently numerous through time, as are perpendicular double platform cores.

TABLE 2 Cores for flakes from Abu Hureyra 1

CORE/ TYPE UNIT 1 UNIT 2 UNIT 3 UNIT 4 NXNXNXNX

Single 38 51. 4 38 73.1 50 62.5 60 70.6

X class X class X class ft class simple 27 71.0 32 84.2 31 62.0 44 73.3 sut-pyr/pyr 8 21.0 4 10.5 17 34.0 15 25.0 subpris/pris 3 7.9 2 5.2 2 4.0 1 1.7

Double 29 39.2 12 23.1 21 26.2 17 20.0

X class X class X class X das? opposing 10 34.5 5 41.7 6 28.6 4 23.5 opposed 11 37.9 2 16.6 4 19.0 5 29.4 perpendicular 5 17.2 5 41.7 11 52.4 7 41.2 adjacent 3 10.3 - - 1 5.9

Multiple 7 9.4 2 3.8 9 11. -3 6 7.1

Other - - - 2 2.3

TOTAL 74 52 80 85

TABLE 3 Cores for Blade/Bladelets from Abu Hureyra 1

CORE/TYPE UNIT 1 UNIT 2 WIT 3 УШ_4 NX NX NX NX

Single 69 32.5 96 34.3 27 21.3 38 33.0

X class X class X class X da?? simple 59 85.5 83 86.4 26 96.3 35 92.1 subpyr/pyr 10 14.5 12 12.5 1 3.7 3 7.9 subpris/pris - 1 1*0 -

Double 48 22.6 83 29.6 40 31.5 18 15.6

X class X class X ÎLeJLèEI opposing 5 10.4 6 7.2 12 30.0 1 5.5 opposed 7 14.6 13 15.7 4 10.0 perpendicular 24 50.0 37 44.6 15 37.5 11 61.1 adjacent 12 25.0 27 32.5 9 22.5 6 33.3

Multiple 92 43.4 96 34.3 53 41.7 55 47.8

Other 3 1.4 5 1.8 7 5.5 4 3.5

TOTAL 212 280 127 113

Data on cores used to manufacture blades and bladelets are presented in Table 3. Here, single platform varieties are most common through time, especially the simple type. There is a noticeable

(10) OLSZEWSKI In : MOORE et al., in preparation. (11) I have reserved the terms subprismatic (subpris) and

prismatic (pris) to refer to cores with a conical shape, rather than to single platform cores in a more general sense, as in CALLEY, 1986. Thus, some may chose to interpret my simple single platform cores as "prismatic".

31



TABLE 4 Comparison of Flakes and Blade/Bladelet Cores from Abu Hureyra 1

CORE UNIT 1 UNIT 2 UNIT 3 UNIT 4 NXNXNXNX

for flakes 212 74.1 280 84.3 127 61.3 115 57.5

for blade/ bladelets 74 25.9 52 15.7 80 38.6 85 42.5

TOTAL 266 332 207 200

TABLE 5 Tool Blank Selection at Abu Hureyra 1

BLANK TYPE UNIT 1 UNIT 2 UNIT 3 UNIT 4 N XN XN XN X

flake 580 52.3 796 55.7 834 55.5 680 39.1 blade 212 19.1 202 14.1 311 20.7 445 25.5 bladelet 281 25.4 391 27.4 331 22.0 576 33.2 other 35 3.2 40 2.8 27 1.8 36 2.1

TOTAL 1108 1429 1503 1737

increase in the frequency of single platform types after Unit 1. Except for Unit 1, double platform cores are most often perpendicular, or of the opposing variety (12).

A comparison of cores for the manufacture of flakes as opposed to blade/bladelets (13) appears in Table 4. While cores for flakes are always more frequent than those for blade/bladelets, their rela- tive frequencies shift through time; cores for flakes decrease after Unit 2, while blade/bladelet cores increase.

Interestingly enough, this decrease in production of flakes also appears to be reflected in the fact that relatively fewer flakes overall are chosen as tool blanks when the earliest three units are compared to Unit 4, the latest in the sequence (Table 5). Thus, tools on blades and bladelets become relatively more common in the latest unit.

Moreover, if the tool classes are examined in terms of the preferred blank types (Table 6), certain constancies, as well as changes, can be observed. Scrapers, burins, notches and denticulates, retouched pieces, multiple tools, and varia are always dominated by flake blanks, although some of these classes do show small percentage increases in blades and bladelets through time. Backed pieces are predominantly on blades, perforators on blades and bladelets, and, by definition, nongeometric and geometric microliths are on bladelet blanks. Trun- cations alternate between flake and blade blanks, with the highest frequency of flake blanks occurring in Unit 2, and blade blanks dominating again by Unit 4.

The percentage data across these categories of debitage, cores and tool blank selection suggest that blade/bladelet production becomes increasingly prominent through time at Abu Hureyra 1. However, the interpretation of this diachronic change is more difficult. On the one hand, it could be that blade/ bladelet technology replaces flake technology, with the overall quantities of blanks manufactured re- maining similar. Alternatively, this change may reflect not so much the replacement of flake technology by increasing use of blade/bladelet technology, as the fact that flake technology is simply less frequently employed through time.

One way to evaluate these alternatives is to assess the data categories against an independent value, rather than against related, and therefore dependent, sets of information, such as percentages. For the purposes of this study, this is accomplished by calculating the density of artifacts in the various categories per cubic meter of excavated sediment (14) for each unit. These density-dependent values (D) therefore more accurately reflect changes through time and can be compared across the units.

Tables 7, 8, and 9 provide the density-dependent data for the categories of debitage, cores and tool blanks at Abu Hureyra 1. Tables 7 and 8 show that flake debitage and cores for the production of flakes are more numerous than their blade and bladelet counterparts for all four of the units. However, an examination of the values for each category (flake, blade, bladelet) suggests that while the number of available blades and bladelets, as well as cores for blades/bladelets, remain relatively constant through time, flakes and cores for the production of flakes decline. This change is most apparent in a comparison of Units 1-3 with Unit 4.

(12) Opposing platforms are located on the same face, while opposed platforms are on different faces of the core. Perpendi- cular platform cores refer to a ninety degree angle between the two platforms, which can be on the same or different faces.

(13) The reader is reminded that these reflect only the end series of removals from the core, and do not necessarily monitor earlier sets of removals. Core-refitting studies, for example, MARKS AND VOLKMAN, 1983, as well as "refitting" through analysis of the exterior surface characteristics of blanks, for example, BAUMLER, 1987, suggest that types of blanks, as well as techniques used to remove those blanks, can change throughout the use-life of a particular core.

(14) The sediments from the levels representing the four chronological units appear to be much the same throughout the sequence. They are generally a light to medium brown in color and are ashy in texture. Patches of orange colored soil, as well as ochre stains and fragments of possible clay floors occur. These have been interpreted by Moore (MOORE et al., in prep.) as the remnants of structural floors that reflect a shift in construction techniques from the pit dwellings found at the base of Unit 1.

32



TABLE 6a Unit 1 Tool Blanks by Tool Class at Abu Hureyra 1

BLANK SCRAPER BURIN PERF BACKED NON6EO 6E0HETRIC NOTCH/DENT TRUNC flULTIPLE HEAVY RET PC VARIA

N X N X N X N II N II N X N X N X N X N X N X N X

flake 186 89.4 « 55.1 18 30.0 21 30.9 173 76.2 30 45.4 10 71.4 2 10.5 79 77.5 18 75.0

blade 11 5.3 30 38.5 12 20.0 47 69.1 48 21.1 34 54.5 4 28.6 21 20.6 3 12.5

bldlet 1 0.5 3 3.8 29 48.3 89 100 153 100 2 0.9 2 1.9 2 8.3

other 10 4.8 2 2.6 1 1.7 4 1.8 17 89.5 1 4.2

TOTAL 208 78 60 68 89 153 227 66 14 19 102 24

TABLE 6b Unit 2 Tool Blanks by Too! Class at Abu Hureyra 1

BLANK SCRAPER BURIN PERF BACKED N0N6E0 6E0HETKIC NOTCH/DENT TRUNCATION MULTIPLE HEAVY RET PC VARIA

N X N X К X N X N X N X N X К X N X N X N X N X

flake 204 95.3 49 65.3 22 28.6 41 32.8 - 253 89.7 50 72.5 9 75.0 3 9.1 143 84.6 22 68.8

blade 5 2.3 20 26.7 26 33.8 84 67.2 21 7.4 19 27.5 2 16.7 21 12.4 4 12.5

bldlet 3 4.0 29 37.6 164 100 177 100 6 2.1 1 8.3 5 2.9 6 18.7

other 5 2.3 3 4.0 2 0.7 - "'0 90.9

TOTAL 214 75 77 125 164 177 282 69 12 33 169 32

TABLE 6c Unit S Tool Blanks by Tool Class at Abu Hureyra 1

BLANK SCRAPER BURIN PERF BACKED N0N6E0 6EDHETR1C NOTCH/PENT TRUNCATION HULTIPLE HEAVY RET PC VARIA

N X N X N X N X N X N X N X N X N X N X N X N X

flake 170 93.4 37 58.7 29 34.1 35 28.0 - 279 79.9 54 54.0 12 70.6 2 11.8 165 73.3 51 83.6

blade 6 3.3 26 41.3 29 34.1 90 72.0 - 53 15.2 46 46.0 3 17.6 48 21.3 10 16.4

bldlet 1 0.5 27 31.7 148 100 131 100 10 2.9 2 11.8 12 5.3

other 5 2.7 - - - 7 2.0 »5 88.2

TOTAL 182 63 85 125 148 131 349 100 17 17 225 61

TABLE 6d Unit 4 Toot Blanks by Tool Class at Abu Hureyra 1

BLANK SCRAPER BURIN PERF BACKED H0N6E0 6E0HETRIC NOTCH/DENT TRUNCATION HULTIPLE HEAVY RET PC VARIA

N X N X N X N X N X N X N X N X N X N X N X N X

flake 130 83.9 62 52.5 18 21.2 37 24.2 190 70.0 41 41.0 9 60.0 159 58.4 34 75.5

blade 17 10.9 52 44.1 38 44.7 116 75.8 63 23.2 59 59.0 4 26.7 88 32.3 8 17.8

bldlet 4 3.4 28 32.9 251 100 256 100 9 3.3 2 13.3 25 9.2 1 2.2

other 8 5.2 1 1.2 9 3.3 16 100 2 4.4

TOTAL 155 118 85 153 251 256 271 100 15 16 1272 45

33



TABLE 7 Debitage by Unit Density at Abu Hureyra 1

CLASS UNIT 1 ПО. 8m3) UNIT 2 (12. 58л3) UNIT 3 (9.6m3) UNIT 4 (22.05m3) ND N Ď ND ND

flake 11,171 1034.3 15,998 1271.7 14,825 1512.7 18,642 845.4 blade 880 81.5 918 72.9 1,238 126.3 1,825 82.8 Ы adelet 683 63.2 782 62.2 940 95.9 978 44.3

TABLE 8 Cores for Flakes and Blade/Bladelets by Unit Density at Abu

Hureyra 1

CORE UNIT 1 (10.8m3) UNIT 2 (12. 58a3) UNIT 3 (9.8ffl3) UNIT 4 (22.05H3) N DN DN D N D

flake 212 19.6 280 22.2 127 12.9 115 5.2 blade/ bl adelet 74 6.8 52 4.1 80 8.2 85 3.8

TABLE 9 Tool Blanks by Unit Density at Abu Hureyra 1

BLANK UNIT 1 (10. 8m3) UNIT 2 (12.58m3) UNIT 3 (9.8m3) MNjT 4 <2?,0?Ц3> NO ND ND ND

flake 580 53.7 796 63.3 834 85.1 680 30.8 blade 212 19.6 202 16.0 311 31.7 445 20.2 bladelet 281 26.0 391 31.1 331 33.8 576 33.2 other 35 3.2 40 3.2 27 2.7 26

Not surprisingly, this shift away from flakes is also reflected in the tool blank information presented in Table 9. The number of tools on blades and on bladelets per cubic meter of sediment is remarka- bly constant throughout the four units (with the Unit 3 blade tools being somewhat anomalous). Tools on flake blanks, on the other hand, decline precipitously in Unit 4. This contrast can also be illustrated by the fact that tools on flakes compared to tools on blades and bladelets, respectively, have ratios ranging from 2.7:1:1.3 in Unit 1, 3.9:1:1.9 in Unit 2, 2.7:1:1.1 in Unit 3, but by Unit 4 this ratio has changed to 1.5:1:1.6. These density-dependent values therefore suggest that it is the manufacture of flakes at Abu Hureyra 1 that becomes less common through time, rather than a replacement of flake technology by an increasing use of blade/bladelet technology.

Summary

Percentage and density-dependent data on debitage, cores and tool blanks from Abu Hureyra 1 suggest that several important chronological changes

occurred at this site. Percentage data indicate that, through time, less primary reduction and less shaping of cores takes place at the site, while shaping and resharpening of tools increases. This may reflect a more intensive use of individual pieces of raw material, rather than a shift to off-site initial core reduction.

Flake debitage, tools on flake blanks, and cores for the production of flakes, based on percentages, always dominate the assemblages from each of the units. However, when density-dependent values are examined, it is clear that the production of flakes, and thus the use of flake blanks for tools, dramati- cally decreases in Unit 4, the latest of the sequence. This suggests that flake technology is "replaced" by blade/bladelet technology at Abu Hureyra 1 in the sense that flake technology is less commonly employed.

Finally, specific tool classes are made on certain blank types consistently through time (eg. scrapers on flake blanks), despite the fact that quantities of flake blanks decline. This indicates that the characteristics of certain types of blanks may have made them maximally effective tools for the particular tasks in which they were employed at Abu Hu- reyra 1. Thus, flake blanks continued to be selected for certain tool classes, rather than replaced by blade/bladelet blanks, despite the reduction in the available number of flake blanks.

COMPARISON WITH TELL MUREYBET DATA

The data from the sequence of units at Abu Hureyra 1 can be compared in part with the information generated in the analyses by Calley(15) for Phase la ("Final Natufian") and for Phase lb ("Epi-Natufian") at Tell Mureybet (16). Phase la consists of Layers 1 and 2, while Phase lb includes Layers 3 and 4(17). These two phases are attributed to the following time ranges : Phase la from 10,500 to 10,300 B.P., and Phase lb 10,300 to 10,200 B.P. Thus, they both appear to be chronologically contemporary with Unit 4 from Abu Hureyra 1. All subsequent observations in this presentation on the debitage, cores and tool blanks from Tell Mureybet are adapted from Calley(18), and will not be presented here in tabular form. The reader is refer- red to the volumes by Calley for full documentation.

Although there are some differences between the layers within each of the two phases at Tell Mureybet, these largely reflect temporal changes in the development of the blade technology of the

(15) CALLEY, 1984; 1986. (16) CALLEY, 1986 : 94-95; CAUVIN J., 1972; 1974; 1977. (17) CALLEY, 1986 : 92-93 (18) CALLEY, 1986.

34



Pre-Pottery Neolithic A and В phases which follow Phases la and lb at Tell Mureybet. For this reason, as well as the fact that two of the layers (Layers 2 and 4) have relatively small sample sizes, the following set of comparisons treats Phase la singly as the combined data from Layers 1 and 2, and Phase lb singly as the combined data from Layers 3 and 4.

Like the debitage classes at Abu Hureyra 1, those from both Phase la and lb at Tell Mureybet exhibit a dominance of flake debitage, with 72.4% flake debitage in Phase la and 74.9% flake debitage in Phase lb (19). The cores from Phase la are primarily single platform types, labelled prismatic for Tell Mureybet, but essentially equivalent to simple single for Abu Hureyra 1. Unlike those from Abu Hu- reyra 1, however, the majority of cores for Phase la appear to be for the production of bladelets (20). Phase lb cores, like those in Phase la, indicate an emphasis on the manufacture of blades and bladelets. Changes between Phases la and lb, noted by Calley (21), include the appearance of bipolar bladelet core technique (opposing platforms in the Abu Hureyra 1 terminology), as well as differences in the preparation of the other cores.

Percentage data for tool blank selection (22) shows that both Phase la and lb at Tell Mureybet have considerably fewer tools on flake blanks than is the case for any of the units at Abu Hureyra 1. Phase la has 20.5 % tools on flake blanks, 29.8 % on blade blanks, 44.6 % on bladelets, and 5.1 % on other blanks. Phase lb shows a reduction in the number of tools on flake blanks, with the following frequencies : 12.7 % flake blanks, 31.0% blade blanks, 54.1 % bladelet blanks, and 2.2 % other. The largest drop in the frequency of tools on flake blanks actually occurs between Layers 1 and 2 of Phase la, from 23.3 % to 12.5 %.

If the debitage and tool blank data from Tell Mureybet Phases la and lb are reworked into density-dependent figures (23), these also show some differences with comparable data from Unit 4 at Abu Hureyra 1. Phase la has a flake to blade/ bladelet ratio of 2.6:1, while Phase lb is relatively similar at 2.9:1 (24). Unit 4 at Abu Hureyra 1 has a ratio of 6.6:1. Tool blank selection is somewhat less disparate between the two sites, with the ratio of flakes to blades to bladelets in Phase la being 1:1.4:2.2, which is relatively similar to the Unit 4 ratio at Abu Hureyra 1 (1.5:1:1.6). However, the Phase lb ratio, 1:2.4:4.3, is quite different from Unit 4 at Abu Hureyra 1.

Summary

The debitage and tool blank selection data from Phase la at Tell Mureybet, both in percentages and density-dependent values, in some instances are roughly similar to Unit 4 from Abu Hureyra 1. The major discrepancies are the flake to blade/bladelet ratio, which is much higher in Unit 4 at Abu Hureyrâ 1, the smaller frequency of tools on flakes at Tell Mureybet, and the greater numbers of cores for the manufacture of flakes at Abu Hureyra 1. In general, the data comparability accords well with the proposed chronological dates for these sets of levels, which fall somewhere between 10,500 and 10,000 B.P., and thus are contemporary with the Late/Final Natufian of the southern Levant.

One reason that the data from Phase la and Unit 4 are not entirely comparable despite the rough contemporaneity of the settlements may be due to differences in site location (on opposite sides of the river) and to slightly different ecological zones. These considerations may affect site function and therefore the tools manufactured at the two sites. Phase la is characterized as having a tool assem- blage with numerous microlithic tools : lunates, triangles, microperforators, perforators, and specialized borers (mèche de foret) (25). As documented above, for Abu Hureyra 1, particular tool classes tend to be made on certain blank types. Thus, scrapers are most often on flake blanks, microliths such as lunates are on bladelet blanks, and perforators are on blade and bladelet blanks. If this observation is not site-specific and therefore confi- ned to Abu Hureyra 1, the greater emphasis on bladelet production in Phase la at Tell Mureybet is explainable by virtue of the fact that the tools produced there are those for which bladelet blanks are preferred.

While the Phase lb assemblages from Tell Mureybet appear chronologically to be contemporary with those from Unit 4 at Abu Hureyra 1, the only similarity in the data is that of the dominance of flake debitage at both sites. The other percentage and density-dependent values clearly show that major differences exist. This suggests that Phase lb at Tell Mureybet may be later than Unit 4 at Abu Hureyra 1, since one possible interpretation of these differences is that Phase lb documents the increasing sophistication of and commitment to the predominantly blade blank production characteristic of the Neolithic. The high percentages of flake debitage in this instance can be seen as the results of the shaping preparation of cores for the manufacture of blades. (19) Ibid. : 375-376, figs. 153-154.

(20) Ibid. : 165. (21) Ibid. : 165-168. (22) Ibid. : 370, figs 148 : 2, 375-376, figs. 153-154. (23) Cubic meters of sediment removed, from CALLEY,

1986 : 374, fig. 152, and data on debitage and tool blanks, from CALLEY, 1986 : 370, fig. 148 : 2, 375-376, figs. 153-154.

(24) Phase la density-dependent values are : flakes - 2046.8, blades/bladelets - 779.3; Phase lb values : flakes - 1920, blades/bladelets « 644.4 (25) CALLEY, 1986 : 93.

35



CONCLUSION

Data generated from the studies of debitage, cores and tool blanks at both Abu Hureyra 1 and Tell Mureybet document the temporal decline in flake production at both sites, However, the reasons for this change at each site appear to be different. The sequence from Abu Hureyra 1 encompasses much of the late Epipaleolithic, and suggests that flake technology is less commonly employed through this period here because of changes in site activities that necessitated tools that were preferentially made on blade and bladelet blanks.

At Tell Mureybet, on the other hand, the sequence spans the interval that incorporates the end of the late Epipaleolithic period and the beginning of the Neolithic. The greatest frequencies of tools here during this portion of the sequence are microlithic, and these are preferentially manufactured on blade/bladelet blanks. Thus, there is a continued emphasis on the production of blades/bladelets through time. At Tell Mureybet, therefore, the change in technology appears to be related to the increased refinement of blade/bladelet manufacture (26) that leads to Neolithic chipped stone assemblages.

Because Abu Hureyra 1 and Tell Mureybet are probably partially contemporaneous, yet have yiel- ded two apparently different assemblages (with the exception of the presence of lunate microliths), is a circumstance that necessitates further consideration. This is a situation not without parallel in other time contexts. Recently, Gilead(27) has observed that one group of Lagaman sites, believed to belong to the Ahmarian Upper Paleolithic tradition, manifests such differences in technology and tool types, that if they had been found in widely separate areas, these sites would be classified as belonging to different chipped stone traditions. The technology employed by prehistoric groups to manufacture chipped stone tools thus profoundly affects the morphology of tools at the type level (28). Such appears to be the case for the late Epipaleolithic at Abu Hureyra 1 and Tell Mureybet. Ultimately, these differences reflect situations of adaptation by prehistoric peoples to different micro-ecological regimes, to activities at such locations, to access to raw material resources, and to temporal changes in activities. These are all aspects that need to be weighed more carefully if current classification schemes and identifications are to be useful in depicting the prehistoric record.

ACKNOWLEDGEMENTS

I would like to thank C. Michael Barton and Andrew Moore, who read earlier drafts of this paper, and made many insightful comments. I, alone, am responsible for the final product. I would also like to thank the Fulbright Program for their financial assistance by sponsering me as a Fulbright Researcher to the Syrian Arab Republic, where I was able to complete a critical portion of the needed research. Finally, and certainly not the least, I would like to thank the following individuals for their per- mission, assistance and time during my stay in the Syrian Arab Republic : Afif Bahnassi, Director-General of the Department of Antiquities, Kasem Touier of the National Museum in Damascus, Wahid Khayata, Director of the National Museum in Aleppo, Hamido Hammade, Curator at the National Museum in Aleppo, and Nazem Jabri of the National Museum in Aleppo.

Deborah I. OLSZEWSKI Department of Anthropology

University of Arizona Tucson, Arizona 85721

(26) Ibid. : 165-168. (27) GILEAD, 1983 : 43, 49-50. (28) MARKS, 1981 : 346.

BIBLIOGRAPHY BAR-YOSEF O. 1981 The Epipaleolithic Complexes in the Southern

Levant. In : CAUVIN J. et SANLAVILLE P., (éd.) Préhistoire du Levant : 389-408. Paris : CNRS.

1983 The Natufian in the Southern Levant. In . YOUNG T.C. Jr., SMITH P.E.L. and MORTENSEN P., (eds.) The Hilly Flanks and Beyond : 1 1-43. Chicago : The Oriental Institute.

BAR-YOSEF O. and VALLA F.R. 1979 L'évolution du Natoufien. Nouvelles suggestions.

Paléorient 5 : 145-152. BAUMLER M.F. 1987 Core Reduction Sequences : An Analysis of Blank

Production in the Middle Paleolithic of Northern Bosnia (Yugoslavia). Ph.D. dissertation. Tucson : University of Arizona. Ann Arbor : University Microfilms International.

BYRD B.F. 1987 Beidha and the Natufian : Variability in Levantine

Settlement and Subsistence. Ph.D. dissertation. Tucson : University of Arizona. Ann Arbor : Uni- versity Microfilms International.

CALLEY S. 1984 Le débitage natoufien de Mureybet : étude prélimi-

naire. Paléorient 10 : 35-48. 1986 Technologie du débitage à Mureybet, Syrie. ix-viiie

millénaire. BAR Int. Ser. 312. Oxford. CAUVIN J. 1972 Nouvelles Fouilles à Tell Mureybet (Syrie)

1971-1972. Rapport Préliminaire. Annales Archéolo- giques Arabes Syriennes 22 : 105-115.

1974 Troisième campagne de fouilles à Tell Mureybet (Syrie) en 1973. Rapport préliminaire. Annales Archéologiques Arabes Syriennes 24 : 47-58.

1977 Les fouilles de Mureybet (1971-1974) et leur signifi- cation pour les origines de la sédentarisation au Proche-Orient. Annual of the American School of Oriental Research 44 : 19-48.

36



CAUVIN M.-C. 1980 Du Natoufien sur l'Euphrate ? In : MARGUERON

J.C1. (éd.) : Le Moyen Euphrate : 1 1-20. Strasbourg : Université des Sciences Humaines de Strasbourg.

GILEAD I. 1983 Upper Paleolithic Occurrences in Sinai and the

Transition to the Epi-Paleolithic in the Southern Levant. Paléorient 9, 1 : 39-53.

HENRY D.O. and SERVELLO A.F. 1974 Compendium of Carbon- 14 Determination Derived

from Near Eastern Prehistoric Deposits. Paléorient 2, 1 : 19-44.

MARKS A.E. 1981 The Upper Paleolithic of the Negev. In : CAUVIN

J. et SANLAVILLE P., (éd.) Préhistoire du Levant : 343-352. Paris : C.N.R.S.

MARKS A.E. and VOLKMAN P.W. 1983 Changing Core Reduction Strategies : A Technolo-

gical Shift from the Middle to the Upper Paleolithic in the Southern Levant. In : TRINKAUS E. (ed.). The Mousterian Legacy : BAR Int. Ser. 164 : 13-33. Oxford.

MOORE A.M.T. 1975 The Excavation of Tell Abu Hureyra : a preliminary

report. Proceedings of the Prehistoric Society 41 : 50-77.

1979 A Pre-Neolithic Farmer's Village on the Euphrates. Scientific American 241, 2 : 62-70.

MOORE A.M.T, GOWLETT J.A.J., HEDGES R.E.M., HILLMAN G.C., LEGGE AJ. and ROWLEY-CONWY P.A. 1986 Radiocarbon Accelerator (AMS) Dates for the

Epipaleolithic Settlement at Abu Hureyra, Syria. Radiocarbon 28, 3 : 1068-1076.

MOORE A.M.T., HILLMAN G.C. and LEGGE A. in prep. Abu Hureyra and the Advent of Agriculture. OLSZEWSKI D.I. 1984 The Early Occupation at Tell Abu Hureyra in the

Context of the Late Epipaleolithic of the Levant. Ph. D. dissertation, Tucson : University of Arizona; Ann Arbor : University Microfilms International.

1986 The North Syrian Late Epipaleolithic : the earliest occupation at Tell Abu Hureyra in the context of the Levantine Late Epipaleolithic. BAR Int. Ser. 309. Oxford.

1988 The North Syrian Late Epipaleolithic and its Relationship to the Natufian Complex. Levant XX : 127-137.

37



Tool Blank Selection, Debitage and Cores from Abu Hureyra 1, Northern Syria

Documents