15 Pages

Neolithic Arrowheads of the Levant : Results and Implications of a Seriation Analysis - article ; n°1 ; vol.15, pg 43-56


Gain access to the library to view online
Learn more


Paléorient - Année 1989 - Volume 15 - Numéro 1 - Pages 43-56
A multi-dimensional seriation analysis of arrowhead assemblages from Neolithic sites in the Levant is presented. An attempt is made to correlate the relative dating obtained by seriation with stratigraphie evidence and available C-14 dates. The implications of the analysis for the study of diffusion processes and sub-regional contacts is also discussed.
Une analyse multidimensionnelle des assemblages de pointes de flèches de sites Néolithiques du Levant est présentée. Une tentative de corrélation des datations relatives obtenues par la stratigraphie et des datations С14 disponibles est faite. Les implications de l'analyse pour l'étude des processus de diffusion et les contacts sub-régionaux sont aussi discutées.
14 pages
Source : Persée ; Ministère de la jeunesse, de l’éducation nationale et de la recherche, Direction de l’enseignement supérieur, Sous-direction des bibliothèques et de la documentation.


Published by
Published 01 January 1989
Reads 31
Language English
Document size 2 MB

Avi Gopher
Neolithic Arrowheads of the Levant : Results and Implications of
a Seriation Analysis
In: Paléorient. 1989, Vol. 15 N°1. pp. 43-56.
A multi-dimensional seriation analysis of arrowhead assemblages from Neolithic sites in the Levant is presented. An attempt is
made to correlate the relative dating obtained by seriation with stratigraphie evidence and available C-14 dates. The implications
of the analysis for the study of diffusion processes and sub-regional contacts is also discussed.
Une analyse multidimensionnelle des assemblages de pointes de flèches de sites Néolithiques du Levant est présentée. Une
tentative de corrélation des datations relatives obtenues par la stratigraphie et des datations С14 disponibles est faite. Les
implications de l'analyse pour l'étude des processus de diffusion et les contacts sub-régionaux sont aussi discutées.
Citer ce document / Cite this document :
Gopher Avi. Neolithic Arrowheads of the Levant : Results and Implications of a Seriation Analysis. In: Paléorient. 1989, Vol. 15
N°1. pp. 43-56.
doi : 10.3406/paleo.1989.4484
vol 15/1 - 1989 PALÉORIENT,
ABSTRACT. - A multi-dimensional seriation analysis of arrowhead assemblages from Neolithic sites in the Levant is presented.
An attempt is made to correlate the relative dating obtained by seriation with stratigraphie evidence and available C-14 dates.
The implications of the analysis for the study of diffusion processes and sub-regional contacts is also discussed.
RESUME. - Une analyse multidimensionnelle des assemblages de pointes de flèches de sites Néolithiques du Levant est présentée.
Une tentative de corrélation des datations relatives obtenues par la stratigraphie et des datations С 14 disponibles est faite. Les
implications de l'analyse pour l'étude des processus de diffusion et les contacts sub-régionaux sont aussi discutées.
INTRODUCTION tion of the general chronology against known str
atigraphie data and the С 14 record.
As the region is geographically large and di
Recent analyses of radiocarbon dating have r verse, and given the levels of information flow and
evealed new problems in the use of this dating me transportation in prehistoric times, a retardation fac
thod. Dendrochronological calibration indicate that tor operated within the general diffusion/contact sys
during some periods in the Holocene radiocarbon tem. This factor can potentially help in calculating
dates tend to expand calendar dates, and in others, rates of diffusion and perhaps status of contacts.
to contract them. That is, calibrated С 14 dates over Such reconstructions, stemming from the analysa span of a particular period may either reduce or is of lithic assemblages, can provide insights into enlarge the apparent calendar years. the understanding of Neolithic communities in the
Given this background, and other known defi Levant, and suggest a model for interpreting other
ciencies of C14 dating (e.g., recovery techniques and aspects of culture.
post-depositional processes), it is apparent that for
some periods, even with dendro-chronological cal
ibrations, archaeologists may have to return to "old-
METHODOLOGY fashioned" relative dating techniques, such as
seriation. Using these methods in sophisticated ways
can allow the construction of a detailed relative chro A. Data - sources and classification nology which can then be compared to a radiometric
record more critically appraised.
The collection of data from lithic assemblages for the Traditional dating methods tax the availability of purposes of this work included the processing of data from suitable large bodies of reliable data. These methods some 100 assemblages from sites throughout the Levant.
need to be applied on elements sensitive to change, These were ordered in several levels of archaeological re
and quickly distributed over well defined geographic liability according to the type and character of the exca
regions. To these ends, flint arrowhead assemblages vation or collection of the assemblages.
were chosen for the analysis and the Mediterranean Level 1 - Assemblages from stratified sites : Despite
Levant as the geographic unit (as defined from the drawbacks, these assemblages are considered highly relia
Middle Euphrates in the North to Southern Israel and ble, to the degree that it was possible to adjust for intrusive
elements. Sinai in the South, including parts of Lebanon, Syria,
Jordan, and Israel). Level 2 - Single stratum sites These are generally
short-lived sites that do not contain remains of earlier or Results of a seriation analysis of Levantine Neol later periods which could disturb and mix the assemblages. ithic arrowhead assemblages indicate that the qual These are considered highly reliable. itative (typological) and quantitative (relative Level 3 — Systematic surface collections : These are frequency of types) changes in these assemblages are assemblages that have been collected using a surface grid chronology dependent. It is also argued that the pro with full control of the context and location of the artifacts
cesses of change in the assemblages are similarly collected. Most cases dealt with here are single stratum
patterned throughout the subregions of the Levant. sites. In some cases these contain more than one component
These processes are also important in the and it is difficult to separate intrusive elements.
43 Colloque Préhistoire Levant II Maison de l'Orient-Lyon
30 mai-4 juin 1988 Editions du CNRS, Paris, 1989 :
4 - Grab samples : These are tool samples colLevel С Seriation
lected in an unsystematic manner. The principal problem
with these assemblages lies in bias by the collector and
Seriation is viewed as a descriptive analytic technique the consequent statistical unreliability of the sample. When
which allows the ordering of comparable units in one dthe assemblage comes from a surface site, the problems
imension so that the placement of each unit reflects its saffecting surface sites exist as well.
imilarity to other units. Level 5 - Varia and problematic assemblages These
Seriation has been used by archaeologists as early as are assemblages of uncertain contexts, insufficient sample Pétrie (4) who analyzed ceramic assemblages from Egyptsize, and other technical problems (collections dispersed, ian tombs. In the early stages the basic assumption for incomplete). Despite difficulties, some of these assem seriation was that each type analyzed has a specific life blages were examined and studied for this work, though span. This assumption was expanded and developed (5) not included in the seriation analysis. such that each type was assumed to have a specific life
span and a specific popularity curve during this span. Each
period can then be characterized not only by the types pre
sent, à la Pétrie, but also by their relative frequencies in
the assemblages. B. Arrowheads and arrowhead typology
The basic assumptions in organizing data in the mat
rices for seriation are : 1. Each tool type had a life span
during which it underwent a progressive increase in its reThe arrowhead was selected as the type for analysis lative quantity in the assemblage until reaching a peak, fofor several reasons : llowed by a progressive decrease in quantity until the final
disappearance of the tool, 2. Chronology is the only factor 1. There are arrowheads in nearly every Neolithic site,
operating and influencing the above described popularity usually in sufficient quantity for processing by quantitative
curve, 3. Each unit (assemblage) examined has a sufficient analysis.
and representative sample of tool types, and 4. The units 2. We can assume, as most researchers do, that the undergoing seriation analysis represent short time segments arrowhead is a tool with a clear function. Thus, it is pos and originate in a limited geographic area containing one sible to concentrate on features that are revealed in the mor cultural system. phology of the arrowhead and how it was prepared. A more sophisticated stage in seriation analysis deals Stylistic differences in the preparation of arrowheads can with multidimensional seriation methods (6). These technibe traced, thus enabling us to distinguish geographical and ques allow a wider chrono-geographic range to be analyzed social units (1). than would be possible using classic one-dimensional se
3. The arrowhead appears in a large variety of types riation precisely because the multi-dimensionality consi
differing from one another in form, size and workmanship ders more than a single factor and can therefore incorporate
and it is given to relatively rapid processes of change. several cultural trends into the analysis. This method is ba
These two features make it suitable for sedation analysis. sed on similarity input matrices and is not limited to uni-
modal popularity curves. The matrix is analyzed using a 4. The appearance of the arrowhead, as well as its di- similarity function that must be correctly chosen to fit the sapperance, are well anchored stratigraphically and are also mathematical analysis and, more importantly, the definition chronologically well defined by C14 dates. of the archaeological meanings of similarity. The multi-d5. Representation of the complete range of compon imensional scaling technique then builds a geometric pre
ents of the analyzed tool group is one of the requirements sentation of the matrix model (the similarity matrix derived
of seriation analysis. It is assumed that the complete range from the original quantitative matrix), so that the geometric
of flint arrowhead types was present in the assemblages (Euclidean) distance between the points appearing in the
examined, thus fulfilling this condition. multi-dimensional space represents the degree of similarity
between the points. The degree of stress applied to transSince the analysis is based on the arrowhead group, form the multi-dimensional matrix into the distance prea detailed typological list is presented in figure 1 The sented on the N-dimensional graph is measured by a types are defined according to morphology and technolo coefficient termed the Coefficient of Alienation (CoA). A gical elements. CoA of 0 indicates absolute identity between the input mat
The detailed definition of types and subtypes was rix and the final graphic result, whereas a CoA of 1 in
bases on my own observations (2) on ca. 15,000 arr dicates complete alienation. A coefficient of 0.15 or less
owheads and on typelists and definitions suggested by indicates good correspondence which can be relied upon.
others (3). However, results of some value can also be obtained when
the coefficient is over 0.2 (7), or when interpretative aThe Harif point was omitted from this list since it rep lternatives to that provided by the S.S.A. are limited or nonresents an Epipaleolithic unit. Its limited spatial and tem existent (8). poral range and its absence from any non-Harifian
assemblage allow analysis on a qualitative level, not re
quiring quantitative methods. (4) PETRIE, 1899 295-301.
(5) BRAINDERD, 1951 301-13; ROBINSON, 1951 293-
(1) GOPHER, 1985. 301.
(2)1985; 53-63. (6) COWGILL, 1968, 1972; KRUSKAL, 1971; KENDALL,
(3) BAR-YOSEF, 1981; CAUVIN M.C. and STORDEUR, 1971; LeBLANC, 1975; DRENNAN, 1976; for review, see MAR-
1978; ECHEGARAY, 1966; BURIAN and FRIEDMAN, 1979; QUARDT, 1978 278-292.
MORTENSEN, 1970; CAUVIN J., 1968; CAUVIN M.C, 1974; (7) TOLEDANO, 1977.
LECHEVALLIER, 1978. (8) GUTTMAN and LEVI, 1980.
44 A9-Herzlia point АЮ-Transverse CM
FIG. 1. - Arrowhead types and subtypes.
45 :
points complete an assemblage in which A18 (a speD. Application
cial type of Byblos point; fig. 1) constitutes a large
majority. We have chosen to use a method of seriation based
on multi-dimensional analysis performed according to the III. Assemblages dominated by a combination
S.S.A. (Smallest Space Analysis) method developed by of Jericho and Byblos points and of the intermediate
Guttman (9) and adapted for use in archaeology by Lin type, A45. These account, at times, for over 90 % goes (10). Remarks on this method of processing can be of the arrowheads. found in Leblanc (11) and Drennan (12). The adaptation
of the method for performance by computer was done by IV. Assemblages in which arrowheads of the
Guttman and Lingoes (13). Different applications were la types A7, A8 and A9, along with their intermediate
ter added by others (14). In Levantine archaeology that t A78 and A89 types, are prominent. Amuq points, echnique has been used by Gopher (15),Goring-Morris (16) A56, and at times, a few transverse arrowheads, apand Sharon et al. (17). pear as well.
V. Two assemblages in which the transverse a
rrowhead forms either an important component or the
RESULTS bulk of the assemblage, and is accompanied by A7,
A8 and A9.
The analysis was conducted in three main
stages : A. Stage 1 - N=67
1. S.S.A. of 67 assemblages of reliability levels
I-IV (dimensions 1-2-3).
Results of the one- and two-dimensional ana2. S.S.A. of portions of the assemblages in order
lyses on the N=67 set were inconclusive, as expected to detail and refine their position within the major (Table 1). Results of the three-dimensional analysis groups obtained from the main analysis (N=44,
are presented in Table 2. N=38).
3. Assemblages of reliability level IV were ex
cluded and separate assemblages from short lived
B. Stage 2 sites were combined (such as Abu Salem; Abu Maadi
III; Nahal Issaron; Nizzana; and others). Fifty assem
blages thus remain to be analyzed. Another sub-pro Groups II and IV contain 46 of the 67 assembcedure of this stage was an S.S.A. analysis on 47 lages, and in order to verify and refine the relative assemblages after excluding the three "outliers" on chronological ordering within these groups, two furthe N=67, N=50 analyses. ther seriation analyses were undertaken :
In spite of minor inconsistencies, the similarity 1. 44 assemblages, principally from Groups II in the results of the five analyses is striking. and III, with a few assemblages from Group IV (as
Five principal groups of assemblages were di defined by the N=67 analysis), and
stinguished on the graphs : 2. 38 assemblages from Groups III and IV
I. Assemblages dominated by the el-Khiam (fig- 2).
point, accompanied by a few Helwan points. One There were 27 assemblages common to both anashould mention the Abu Maadi I assemblages in lyses. This allowed the comparison of the results which el-Khiam points are accompanied by a specif from the two analyses and provided double verificaic type called the Abu Maadi point (18) which is tion of the position of these assemblages. similar to the Harif point, and is absent from all
other assemblages.
II. Assemblages where the main arrowhead type
C. Stage 3 is the Helwan point, accompanied by Jericho and By-
blos points in considerable quantities. Notice should
be made of the Mureybet XII-XVII assemblage in The exclusion of low reliability assemblages a
which Helwan points accompanied by el-Khiam llowed an even clearer analysis. The results were still
very similar to previous analyses with respect to ge
(9) GUTTMAN, 1968 469-506. neral groups and the specific placing of assemblages (10) LINGOES, 1970 277-97, 1973. within the relative chronology (fig. 3). (11) LeBLANC, 1975 22-38.
(12) DRENNAN, 1976 290-302. An attempt was made to achieve more accuracy
(13) LINGOES, 1966. by excluding three assemblages bearing types of a(14) TOLEDANO, 1977. rrowheads that do not or rarely appear in other a(15) GOPHER, 1985. ssemblages (Al, A18, A10). Both Stage III analyses (16) GORING-MORRIS, 1987.
had low Coefficients of Alienation, actually presen(17) SHARON et al., 1987.
(18) BAR-YOSEF, 1981. ting a classic one-dimensional seriation (Table 1).
46 TABLE 1 In other assemblages, such as Mureybet I-VIII,
Coefficients of Alienation in different Smallest Space Analyses Mureybet IX-XI (Cauvin's Upper Layer II, Lower III
employed in the sedation respectively) and Nahal Oren, el-Khiam points are
accompanied by varying quantities of Helwan points. Stage II Stage III Stage I S S A The Abu Maadi I assemblage is exceptional in the N = 67 N = 44 = 38 N = 50 N = 47 N
composition of arrowhead types in that it contains Dimensionality a majority of Abu Maadi points, accompanied by ma.124 .121 1 417 .164 .244 ny el-Khiam points. 2 .224 .128 .096 .016 .105
3 .142 .117 .069 .001 .076 In the Mureybet layers, the el-Khiam point unreduction in the number of assemblages (N) or variables 1) Usually, derwent an increase in popularity (beginning in Cauv(types of arrowheads) causes a reduction in CoA. in's layer Ib) to reach a peak in Layer II and then 2) Elimination of assemblages from reliability level IV improves the a decline until disappearance. The Helwan point quality of the analysis (see N = 44). This phenomenon does jnot stand
out in the N = 47 analysis due to the elimination of all the outliers. joins the arrowhead assemblage at a very early stage
3) No analysis was carried out if the CoA was higher than 0.16. (Van Loon's I-VIII or Cauvin's Upper II) and under
goes a similar process of progressive increase and
afterwards decreasing popularity. It is thus clear that
it is the second type in the arrowhead succession in DISCUSSION
the assemblages from the northern Levant. Van
Loon's IX-XI have been attributed to Group I due
to their stratigraphie position and the fact that el- Combining all five analyses in a general scheme
Khiam points comprise almost 2/3 of the arrowheads. still shows the same general groupings as stated
above. It is beyond the scope of this paper to discuss A generally similar process can be followed at each assemblage individually; however, general ten Nahal Oren despite stratigraphic/contextual prodencies can be noted in relation to each of the main blems (21). groups.
Farther south in the Negev, the Nahal Lav an 108
arrowhead assemblage is dominated by el-Khiam
points with straight bases or knob or button-shape Group I
tangs. The regular bilaterally notched type is almost
absent. A significant feature in this assemblage is the Group I included nine assemblages from eight appearance of Harif points (22), microliths, inclusites. The sites of this group are dispersed over a ding lunates, and the use of the microburin techni
wide and diverse area from the Middle Euphrates que. These suggest affinities to a Harifian-like
to Southern Sinai. Arrowheads usually comprise a re industry.
latively small percentage of the tools in the larger
The southernmost assemblage, Abu Maadi I, sites (e.g., Mureybet, Jericho, Netiv Hagdud) and a
contains a significant component of microliths, ilarger one in smaller sites (e.g., Salibiya IX, N. La-
ncluding many geometries, especially lunates and van 108, Abu Maadi I). The arrowhead group
triangles. The Abu Maadi points show similarities to consists chiefly of a single type, the el-Khiam point.
Harif points, especially the subtype lacking the knob- It is therefore difficult to arrange the sites in a chro
like tang. Similarities may be observed in other feanological order derived from the seriation analysis.
tures as well (raw materials, bell-shaped In only a few cases do other types join the el-Khiam
pestles, etc.) and the general similarity of Abu Maadpoints to support a chronological ordering, together
i I to the Harifian is thus to be noted. with С 14 dates (insofar as they exist and are not iso
lated dates) and stratigraphie evidence. Stratigraphically, Group I assemblages originate
The Jericho PPNA, Salibiya IX, Netiv Hagdud, from layers above Natufian layers (Mureybet, Jeri
and Poleg 18-M assemblages contain only el-Khiam cho, Nahal Oren) and beneath PPNB layers (Mureyb
points. However, one should note that these also in et, Jericho, Nahal Oren, Abu Maadi I).
clude a range of subtypes which have not been trea Many clusters of radiocarbon dates are available ted here as denominators for internal relative from Group I sites. It is possible to reconstruct the chronology (19). These subtypes include the classic following radiometric chronological scheme : bilaterally notched el-Khiam points (fig. 1:A2-I), the
"tanged" (fig. 1:A2-III) and the straight based or The Mureybet series suggests a date of
knob-base subtypes (fig. 1:A2-IV), currently being ca. 8200 b.c. for the appearance of el-Khiam points
analyzed (20). This more detailed analysis may en (layer Ib), a peak in its popularity 8200-8100 b.c.
able refinement of the analysis for relative dating wi and disapperance in the first third of the 8th millen
thin Group I(PPNA) assemblages. nium b.c.
(19) GOPHER 1985 : 54-5. (21) GOPHER, 1985 : 224-25.
(20) NADEL D., pers. comm. (22) pers. observation; GORING-MORRIS, 1987 : 332.
47 :
The Jericho series shows three clusters of dates points, and various other points like the Harif Point
within the PPNA layer, one ranging 8350 to (N. Lavan 108, various Harifian sites) and the Abu
8230 b.c. (4 dates), the second 7825-7610 b.c. (4 Maadi point (Abu Maadi I). These all lack bifacials
dates), and the third with one subcluster 7480- and reaping knives.
7430 b.c. (4 dates) and one 7370-
7250 b.c. (5 dates). The consistency of the dates
with the stratigraphy is far from satisfactory. The Group II
wide gap between the earliest PPNA cluster and the
later ones can be interpreted as a real gap in the site Group II includes nine assemblages from eight history, or, perhaps, a reflection of the various tech sites throughout the Levant from the Negev to the nical problems in the dating method itself, as alluded Middle Euphrates. In most sites bipolar cores were to at the beginning of this paper. It is worth noting used for the production of blades and this charactthat each cluster of dates is uniform and short ran eristic (with a full range of features) is known as ged, even though the analyses were undertaken in early as Mureybet XII-XVII (28) and Tell Aswad la. different laboratories. The samples from the latest
The arrowhead assemblages of Group II are cluster originate in late stratigraphie phases in the
composed of several types. The most prominent is Jericho sequence, as opposed to all other clusters,
the Helwan point, which made its initial appearance and it has been suggested that these do not belong
in the earlier Group I assemblages. Byblos and Jeto the PPNA occupation layers (23).
richo points constitute the remainder. Some of the The 12 Netiv Hagdud dates show a range of Group II assemblages have el-Khiam points, at times 7830 to 7650 (excluding one earlier and one later in rather significant numbers (e.g., Mureybet Midddates). The four nearby Gilgal I dates are somewhat le III), but usually only in very low frequencies earlier, ca. 7 900 b.c. (24). Four more dates from (e.g., Nahal Boker, Michmoret 26). Gesher, 100 km north of Jericho (25) fall late in the
Helwan points are a variable group and diffesecond half of the 9th millennium and beginning of
rences in morphology and workmanship, as well as the 8th millennium b.c.
in blanks, can be discerned between subregions of In the south, at Abu Maadi I, a series of seven the Levant (29). It is important to emphasize that in dates (26) date the lower layers to 8020-7840 (ex all analyses the Helwan point bearing assemblages cluding one earlier date). It is of importance in this were grouped together and clustered in the earliest context to note that twenty С 14 dates from Negev part of the PPNB sequence. Harifian sites (27) range from 8 600 to 8 050 b.c.
Analyzing the stratigraphie position of the aand may suggest a retardation factor explaining the
ssemblages shows that they are positioned above the later South Sinai derivatives of this entity (i.e., Abu
Group I assemblages (Mureybet, Nahal Oren, Abu Maadi I).
Salem [above the Harifian]) and under later PPNB Summing up this section, we may divide the assemblages (Mureybet, Nahal Oren, Tell Aswad). Group I assemblage as follows :
Some points should be noted concerning the 1. Assemblages containing el-Khiam points and composition of Group II assemblages. microliths (including geometries) as prominent ele Mureybet Middle III includes groups of el- ments. These lack bifacials and reaping knives (Sa-
Khiam points and Helwan points, but the largest libiya IX, Poleg 18-M, el-Khiam 3-4).
component is a large coarse type of Byblos point, 2. Assemblages with a small microlithic element here called A18 (30). The attribution of this assemor none at all, which contain bifacials and reaping blage to Group II stems from the fact that the el- knives (Beit Taamir knives), such as Mureybet Ib, Khiam and Helwan points appear in it and constitute Netiv Hagdud, and Jericho. 1/3 of the arrowhead assemblage, and from its str
3. Assemblages lacking a microlithic element, atigraphie position and radiometric dating.
that contain an arrowhead group composed of el- In the Central and Northern Levant, Byblos and
Khiam and Helwan points (e.g., Mureybet upper Byblos-like points join the Helwan points. Jericho layer II, lower III, possibly Nahal Oren). Bifacials points are rare to absent. In the Southern Levant, Je
(including the erminette) and reaping knives (inclu richo points are an important component in Group II ding the Beit Taamir) also appear in these assemb assemblages (Nahal Oren, Michmoret 26, Abu Salages. lem) and remain so through most of the PPNB. This
4. Southern assemblages containing microliths typological variation may reflect local adaptations,
(including geometries), small numbers of el-Khiam as for example, the erminette of the Northern Levant
probably represents a similar phenomenon for the
axe/adze group. (23) GOPHER, 1985 300-304.
(24) NOY, CARMI and VOGEL, pers. comm.
(25) GARFINKEL, pers. comm. (28) CAUVIN J., 1978 : 32; fig. 13.
(26) BAR-YOSEF, pers. (29) GOPHER, 1985 228-229.
(27) GORING-MORRIS, 1987 : 449-452. (30) Fig. 1; also GOPHER 1985 : 58.
48 :
\ \ ч v
30 ,*o
27. - :
.32J.3.7 » •
50 too
- The two-dimensional representation of the N=38 S.S.A. FIG. 2.
Item 27. Haluza Dunes 85 Item 1. Munhata 6-3 Item 14. Abu Maadi III (upper)
15. Qadesh Barnea 3 28. 94 2. 2 (clean)
3. Munhata 2 (general) 16. 31 29. N. Sekher 81 -A
4. Jericho PPNB (early) 17. Mushabi VI 30. Abu Gosh (upper)
31. Abu (lower) 5. (late) 18. N. Issaron (combined)
6. Jericho PN 19. Beidha VI 32. Yiftahel
7. W. Tbeik 20. V 33. Sha'ar Hagolan
8. Ujrat el-Mehed 21. Beidha IV 34. Givat Haparsah
35. Herzliya 9. W. Jibba I 22. III
10. W. II 23. Beidha II 36. Nizzanim
11. W. Jibba IIA 24. I 37. Qatif (Y-3)
38. Byblos Neo. Anc. 12. Abu Maadi III (lower) 25. Ramat Matred V
13. Abu III (middle) 26. VI
The discussion of the absolute dating of the Group III
Group II assemblages is marked by the paucity of
С 14 dates, especially clusters of many dates from
Over 30 assemblages from 22 sites in the Southe same site. The Mureybet series shows that the
thern Levant ranging from the Galilee to Southern Helwan point made its appearance in Upper Layer II, Sinai are included in this group. The absence of i.e., ca. 8100 b.c., or perhaps somewhat later, and northern assemblages is conspicuous and due to inacreached a popularity peak around 8000 b.c.. Farther cessibility. Therefore, these will be discussed only south, in the Damascus Basin, the Tell Aswad la on the basis of the published data. dates suggest a range from ca. 7900 to
7600 b.c. (31) for the popularity peak of this type. Technologically, in most of these assemblages
In the southern Levant, dates from Sefunim Cave bipolar naviform cores are used for the production
(not included in the analysis due to small sample of long straight blades. These cores appear in varying
size) point to a mid-8th millennium b.c. date for Hel quantities in each of the assemblages (33), though
wan dominance (32) and to the end of the 8th they are not always dominant. The blade/flake ratio
millennium and beginning of the 7th for the disap in the debitage is not always in favor of blades. The
pearance stage. Depicting these data on a space/time "blade character" often attributed to PPNB assem
graph (fig. 4) enables the reconstruction of the dif blages is, in effect, a general characterization of the
fusion process of the Helwan point from its area of ability to produce long straight blades, and the ten
origin on the Middle Euphrates to the southern parts dency to select them for the fashioning of tools (es
of Israel, and the calculation of its retardation factor pecially arrowheads and sickle blades).
(fig. 5). The arrowhead group is prominent in its relative
frequency, reaching up to 1/3 of the tools. I will not
(33) e.g., CROWFOOT-PAYNE, 1983 : 666; MORTENSEN, (31) CAUVIN M.C., 1974.
1970 9; YIZRAELI, 1961 : 41; GOPHER, 1981 70. (32) RONEN, 1984.
49 58.0
5О5 51.0 51.5 52.0 52.5 53.0 53.5 54.0
FIG. 3. - The two-dimensional representation of the N=50 S.S.A.
1. Munhata 6-3 18. Ein Qadeis 35. Nahal Sekher 81-A
2. 2 19. Nahal Issaron (C) 36. Abou Gosh (upper)
3. Jericho PPNA 20. Reuel 37. (lower)
38. Yiftahel 4. Jericho PPNB (early) 21. Beidha VI
5. (late) 22. V 39. Sha'ar Hagolan
6. Jericho PN 23. Beidha IV 40. Salibiya IX
7. Nahal Oren PPNB 24. III 41. Netiv Hagdud
8. PPNA 42. Givat Haparsah 25. Beidha II
9. Wadi Tbeik 26. I 43. Poleg 18-M
10. Ujrat el-Mehed 27. Nahal Boker 44. Michmoret 26
11. Wadi Jibba I 28. Hemar 45. Herzliya
12. II 29. Nahal Divshon 46. 26A
47. Mureybet I-VIII 13. Wadi Jibba Ha 30. Abu Salem
14. Abu Maadi I 31. Kvish Harif 48. IX-XI
15. Abu III 32. Nizzana 49. Mureybet XII-XVII
16. Gebel Rubsha 50. Byblos Néolithique 33. Nahal Lavan 109
17. Kadesh Barnea 3 34. 108
enlarge on the description of arrowhead technology, Examination of the composition of arrowhead
but only note that there is a perceptible difference assemblages of Group III shows that in only a few
in size and other stylistic elements between the a cases does one type dominate. The assemblages are
rrowheads from different subregions. usually composed of several arrowhead types, and it
is possible to distinguish clear trends of change in Stratigraphy of Group III assemblages is meag
the relative frequencies of each type. er. Only a few sites contribute information. Under
Group III assemblages, one finds either Group II or Two main subgroups of arrowhead compositions I assemblages (e.g., Jericho, Abu Madi I, Bei are distinguishable : an early one in which Jericho, dha). Group III assemblages are stratified under pot Byblos, and the intermediate A45 types are domintery bearing layers of the 6-5th millennia b.c., when ant, and a later one, where Byblos and Amuq they have layers above them (e.g., Jericho, Munhata, points (34) (and the A56) are dominant. Abou Gosh).
(34) Amuq points are here mainly "leaf-shaped" points and do
not usually follow the original definition by J. Cauvin (1968).
50 :
/ \
/ /\ \
/TP AIN QADEIS L) \ \ \ JERICHO (BASE) PPNB / < / /I I-, Г \ I/' \ \ \\ V / '/. V / ' sefuniv. VAN IOON JCAUVIN jk NAHAL ' \ \ cav: ? | L AVAN /' _J 109 i
ORE nahal MICHMORET N alJ iU 26.26 A \\ _
□' NAHAL / OREN I ->
\ I / \ / JZ-XL \
I УЖ 1_LJ/ П.
FIG. 4. - The Helwan point diffusion process.
Excluding the reliability level IV assemblages, of the Byblos point is joined by Amuq points, and
the above stated general trends of change in their number steadily increases. It seems that the By
Group III arrowhead assemblage composition are blos point and probably the Amuq point as well are
summarized in Table 3. With only a few exceptions, of a northern origin and were diffused southwards.
It is of importance to note that the Jericho point is the general picture is clear and can be graphically
expressed for subgroup Ilia in figure 6. This trend almost completely missing from the Central and
is continued in the assemblages of subgroup Illb Northern Levant, although it exists as a rare type in
(fig. 7). Analyzing these data, one can conclude that some Damascus basin assemblages (35).
the Jericho point was at its peak popularity earlier Some fifty C14 dates are available for the than the Byblos point. The interplay of Jericho and Group III assemblages (30 from Jericho and Beidha), Byblos point frequencies is such that while Jericho placing them between the last third of the 8th milpoints "pull" an assemblage towards an earlier pla lennium b.c. and the close of the 7th millennium cement as their quantity increases, the Byblos points b.c. (36). take the part of a stabilizing intermediary in the sys
tem, though their relative frequency ranges from 1/3-
2/3 of the arrowheads. Thus, the relative
Group IV chronological position of the Group III assemblages
can actually be determined by the ratio of Jericho
points (+ A45) to Amuq points (+ A56).
Group IV includes assemblages from 13 sites,
The Group Illb assemblages thus show a rising these are dispersed in the Central and Southern Le
percentage of Amuq (+ A56) points and a declining vant from Byblos to the Gulf of Suez. of Jericho (+ A45) (Table 3). Bipolar cores are still in use in some of these
assemblages, but are rare. There is a decrease in proIn the Central and Northern Levant, the changes
duction of blades (compared to Group III) and in in arrowhead assemblages are somewhat different.
Coarse, heavy, large Byblos points replace the Hel
wan point as early as the first half of the 8th mil (35) e.g., CAUVIN M.C., 1974 494.
lennium b.c. In later stages, a more refined version (36) For lists of C-14 dates, see GOPHER, 1985 Appendix С