Journal of Archaeological Science (2001) 28, 127–141doi:10.1006/jasc.1999.0545, available online at http://www.idealibrary.com on

From Lithics to Spatial and Social Organization: Interpretingthe Lithic Distribution and Raw Material Composition at theFinal Palaeolithic Site of Kettig (Central Rhineland, Germany)

Michael Baales*

Forschungsbereich Altsteinzeit, Romisch-Germanisches Zentralmuseum Mainz, Schloss Monrepos,D-56567 Neuwied, Germany

(Received 10 October 1999, revised manuscript accepted 8 December 1999)

During the Final Palaeolithic Federmessergruppen period, a wide range of siliceous raw materials was used in theCentral Rhineland (Germany). Beside materials from the immediate region, exogenous resources such as flint from upto 100 km away in the Meuse drainage area and the moraines of the Saalian glaciation were exploited. At Kettig, thespatial distribution of different lithic raw materials and analysis of several categories of burnt finds allow thecharacterization of the settled area in space and time. While adequate, good-quality raw material was available inthe Central Rhineland, the regular use of exogenous raw materials at Central Rhineland sites also demonstrates thenecessity of maintaining regular social contacts with other groups in neighbouring regions. � 2001 Academic Press

Keywords: FINAL PALAEOLITHIC, FEDERMESSERGRUPPEN, SPATIAL ANALYSIS, RAW MATERIALECONOMY, SOCIAL CONTACTS, WESTERN GERMANY.

*E-mail: mbaales.Monrepos@rz-online.de

Introduction

A t Kettig, in the Neuwied Basin (CentralRhineland, Western Germany), it was possibleto investigate a Final Palaeolithic Federmesser-

gruppen site which is of importance for two majorreasons. One of these is given by artefacts of red deerantler of types previously unknown for the Feder-messergruppen; the other is the possibility of a detailedspatial analysis of the site, which gives new insightsinto the internal organization and division of arelatively large habitation area of the period some13·2 ky cal-.

Furthermore, the overall spectrum of lithic rawmaterial observed at Federmessergruppen sites in theCentral Rhineland permits a final statement of generalimportance for human behaviour. While materialssuitable for producing lithic tools were available in theregion, an important amount of material was broughtto the sites from exogenous sources up to 100 kmand more distant. This is interpreted as reflectingregular social contacts with people in neighbouringregions, which were of great importance for smallhunter–gatherer groups in a largely uninhabitedlandscape.

During the Pleistocene, the Central Rhineland and,in particular, the Neuwied Basin to the northwest ofKoblenz, was repeatedly the scene of major volcaniceruptions. The most recent eruption occurred towards

1270305–4403/01/020217+15 $35.00/0

the end of the late glacial Allerød interstadial c. 12·9 kycal- in what is today the Laacher See caldera (Streetet al., 1994; Joris & Weninger, 2000). The massivePlinian eruption covered the late glacial CentralRhineland landscape (Ikinger, 1996) with layers of ashand pumice (Laacher See tephra=LST), often severalmetres thick (van den Bogaard & Schmincke, 1985),and so protected archaeological sites from erosion anddestruction. Pumice quarrying, which began in themiddle of the 19th century and has been carried out onan industrial scale since the Second World War, hasrepeatedly uncovered well-preserved sites, especially ofAllerød age.

The conserving effect of the volcanic sediments hasprotected not only sites of human settlement but alsonumerous botanical localities with preserved trees,imprints of leaves in ash layers of the LST, etc. All thisallows us to draw a very detailed picture of thevegetation and environment of the final Allerød(Baales & Street, 1996; Baales et al., 1998a; Street &Baales, 1997). Generally, the Central Rhineland was alightly forested region with aspen, birch, willow nearthe rivers and some pine on higher, drier grounds. Theregular presence of horse at Allerød sites may alsoindicate more open areas, perhaps at higher elevationssurrounding the Neuwied Basin.

The first archaeological investigation beneaththe LST was carried out in 1883 by HermannSchaaffhausen at the Martinsberg in Andernach. Morerecent excavations have been carried out there in the

� 2001 Academic Press

128 M. Baales

Table 1. The Allerød fauna from Kettig (without molluscs identified byMania)

Species NISP MNI

PerissodactylaHorse Equus sp. 17 3

ArtiodactylaLarge bovid (aurochs?) Bos primigenius? 17 2Chamois? Rupicapra rupicapra 1 1Red deer Cervus elaphus 259 8Roe deer Capreolus capreolus 24 3

CarnivoraBear Ursus arctos 5Wolf Canis lupus 2 1Red fox Vulpes vulpes 22 1 (2?)Weasel Mustela nivalis 1 1Marten Martes sp. 6 1

RodentiaBeaver Castor fiber 29 2Yellow-necked

field-mouse Apodemus flavicollis 3 1Large Pleistocene mole Talpa europaea magna 6 2Common hamster Cricetus cricetus 1 1Bank vole Clethrionomys glareolus 120 18Water vole Arvicola terrestris 40 6Root vole Microtus oeconomus 9 5Field vole Microtus agrestis 8 3Voles Microtus arvalis/agrestis 22 14Voles Microtus sp. 83 —

InsectivoraCommon shrew Sorex ex. gr. araneus 32 11Pygmy shrew Sorex cf. minutus 2 1Water shrew Neomys fodiens 1 1

ChiropteraNoctule Nyctalus sp. 3 1

AvesTit Parus sp. 1 1

PiscesPike Esox lucius 5 1Cyprinids Leuciscus sp. >100 —

Identifications: Baales (1998) and Kalthoff (1998) and Krey (inBaales, in press). NISP: Number of identified specimen; MNI:Minimum number of individuals. The animals certainly exploited byman are red and roe deer, horse, a large bovid (aurochs?), a caprid (ifidentification is correct) and beaver. Possibly exploited were brownbear, wolf, red fox, weasel, marten and the fishes, which are onlyrepresented by pharyngeal teeth.

1980s and from 1994 to 1996 (Baales & Street, 1998).Further important late glacial sites have been dis-covered since, especially during the second half of the20th century, among them the well-known Late UpperPalaeolithic Magdalenian site (c. 15·5 ky cal-) ofGonnersdorf (see Bosinski et al., 1995). Particularlyimportant are four major (late) Allerød settlementsof the Final Palaeolithic Federmessergruppen(Schwabedissen, 1954) (Figure 1). Urbar, discovered in1966, lies to the northeast of Koblenz (Baales et al.,1998b). At Andernach-Martinsberg, an extensive Fed-ermessergruppen horizon (Veil, 1982; Stapert & Street,1997) overlies a Magdalenian settlement equivalentto Gonnersdorf. Niederbieber, a suburb of Neuwied,covers an area of some 10,000 m2 and is the largest siteknown from this period in the Central Rhineland

(Bolus, 1992; Baales, 1998; Baales & Street, 1998).Kettig (District Mayen-Koblenz), which is the subjectof this paper, lies south of the Rhine and was investi-gated over an area of 242 m2 during 6 monthsfollowing its discovery in April 1993 (Figure 2).

The Kettig SiteDuring the Allerød, the Kettig site occupied a prom-ontory of land (the Rhine middle terrace) raised some-what above the banks of the Rhine. A channel at thefoot of the promontory was either permanently water-filled or at least occasionally flooded. A west–eastoriented trial trench established that archaeologicalmaterial was located not only on the promontory itself,but also extended downslope into the Rhine valley-bottom (see Figure 4). There is one conventional14C-date for the site of 11·314�50 (Hd-18123; B.Kromer, pers. comm., Heidelberg), carried out onbulked bones of red deer size. The data was calibrated(‘‘converted’’ following Joris & Weninger, 2000) toc. 13·2 ky cal-.

Unique to Kettig are two artefacts of red deer antler.One of these is a barbed point which morphologicallyand technologically resembles Magdalenian specimensbut is defined by J. Tinnes (in Baales, in press) asan independent ‘‘Kettig type’’. The other artefact is anantler hammer (percuteur), a form of tool otherwiseunknown from a Federmessergruppen context. Despitethe protective covering of LST, Final Palaeolithicartefacts of organic materials have only been rarelyfound in the Central Rhineland (Baales, 1994, 1997,1999; Bolus, 1992; Baales & Street, 1996). Thisis a striking difference to the preceding UpperMagdalenian dated at Gonnersdorf and Andernach toaround 15·5 ky cal-.

The list of large mammal species is dominated by reddeer (Cervus elaphus), which is represented by a mini-mum of eight individuals. The remains of other largemammal species are much less common and representfewer individuals (Table 1).

A number of remains of red and roe deer, especiallyteeth (particularly the analysis of the cement structureof a red deer tooth by Kierdorf [in Baales, in press])allowed the identification of the season of occupationof the site as late summer or early autumn.

The Kettig lithic assemblage is typical of theRhineland Allerød sites. It is dominated by more than100 short end scrapers, followed by projectile points(including both classic Federmesser points, other typesof backed pieces and simple microliths), burins, trun-cated and laterally retouched pieces, borers and splin-tered pieces (pieces esquillees). The spectrum of rawmaterials is generally similar to that of other CentralRhineland sites and is particularly close to that fromAndernach-Martinsberg (Figure 3; see Floss, 1994).The assemblage is made not only of regionally avail-able materials, such as Tertiary limno-quartzite and

From Lithics to Spatial and Social Organization 129

0 3 km

N

BROHL

LaacherSee

10

2

911

13

1

NE

TTE

12 8

7

65

MO

SELL

E

3

4

RHINE

WIE

D

Neuwied

Koblenz

200

300

100200

100

300

400

500

400

300

200

200100

300

100

200100

300

0 200

km

Figure 1. Location of Central Rhineland (Neuwied Basin) Federmessergruppen sites. Settlement (large �): 1—Kettig; 2—Andernach-Martinsberg; 3—Niederbieber; 4—Urbar. Isolated hearth (only two were associated with artefacts, �): 5—Bassenheim; 6—Kobern; 7—Polch;8—Ochtendung; 9—Meisenheim 3; 10—Nickenich. Single finds (�): 11—Miesenheim 2; 12—Thur. Human skeleton (�): 13—Weissenthurm(see Baales & Street, 1996).

Figure 2. Kettig. View of the site from the south. In the background, the Mulheim-Karlich nuclear power plant (some 1·6 km distant) marksthe south bank of the Rhine at the present day.

130 M. Baales

0

Material

Per

cen

tage

pre

sen

t

10

20

30

KS

64

4558

VT

193

8

TQ

1279

9112

VK

7 0

CH

245

337

FW. Euro(Meuse)

Local Eifel+

“Meuse egg”

9319173

KOO

57 0

FBaltic/

Tétange

17870

KA

0 1

LIM

0 2

Q

0 1

VTO

0 1

A/J

0 2

?

0 33

40

Figure 3. Lithic raw material composition of the Kettig assemblage in comparison with the Federmessergruppen assemblage fromAndernach-Martinsberg (1980s excavations). Leaving aside several smaller categories of material, only siliceous slate is distinctly more commonat the Kettig site (numbers above the bars indicate numbers of artefacts). VT: siliceous tuff; KS: siliceous slate; TQ: Tertiary limno-quartzite;VK: siliceous chalk; CH: chalcedony; F: flint; KOO: siliceous oolite; KA: carneol; LIM: limonite; Q: quartz; VTO: siliceous clay stone;A/J: agate/jasper; ?: unidentified. Andernach (�), N=2·793; Kettig ( ), N=24·098.

Joining the Dots? Site Formation at KettigShown by the Lithic ArtefactsPlotting lithic artefacts at a Stone Age site commonlyreveals horizontal clusters, generally interpreted asmaterial left over by the settlement process (mainly inthe form of waste products or debitage). At Kettig, theconserving effect of pumice of the LST resting abovethe find horizon has preserved such evidence for theactivities of the Federmessergruppen inhabitants in amore or less undisturbed condition. Although mappinglithic artefacts is one of the basic methods used in thereconstruction of occupational histories of Palaeolithicand Mesolithic sites, there is almost no consensus as tohow to do this or of how to represent them on a map,a problem discussed in detail by Cziesla (1990). Here Ifollow Cziesla’s proposal in using an ‘‘equidistance’’level for interpreting artefact distributions, and thefollowing maps are based on this procedure.

A ‘‘equidistance’’ plot of the distribution of theentire flaked stone assemblage (Figure 4(a)) shows amajor dense concentration of artefacts at the southwestof the site. Further concentrations with fewer artefactscan be recognized at the northeast, and a still smalleraccumulation is also visible to the northwest. In con-

trast, the rest of the site presents an undifferentiatedpicture. This is due to the fact that any smallerconcentrations present are ‘‘swallowed up’’ by thedisproportionally large number of artefacts of thesouthwestern concentration.

In order to obtain a better resolution for potentiallysmaller lithic concentrations, a second plot was pro-duced (Figure 4(b)) by dividing the site into severalindependent sectors (areas A–F) and using the ‘‘equi-distance method’’ independently for each of these. Thedensest concentration to the southwest (area E) andthe northeastern part of the site characterized by threedistinct concentrations (area B) are again clearly vis-ible. In contrast with Figure 4(a), several previouslyinvisible concentrations of artefacts now appear. Tothe south (area D), a smaller concentration is visible,which ‘‘fades out’’ to the northeast and was unfortu-nately partially destroyed by LST quarrying. A furtherloose scatter of finds to the west (area F) shows noclear structuring and is recognizable by only a re-stricted number of artefacts. It is noticeable that thedensest concentration of this area lies at the boundaryto area E. This is an unusual distribution which isrepeated by other categories of material (e.g. burins,burned artefacts). Another small concentration with afew retouched forms is found down the eastern slope ofthe promontory (area C), a situation hard to interpretdue to limited excavation in this area. The interpret-ation of the northernmost part of the site (area A) isnow particularly clear and a major concentrationof artefacts can be recognized. Unfortunately, thesouthern margin of this area was truncated byquarrying.

Figure 4(b) characterizes Kettig as a site comprisingvarious small concentrations of greatly varying density.

siliceous slate (Kieselschiefer or lydite), but alsoincludes chalcedony from Bonn-Muffendorf, some40 km to the north (Floss, 1992), and large quantitiesof several types of Cretaceous flint, particularly fromthe Meuse region c. 100 km to the northwest of thesite (see Baales & Street, 1996). Spatial plotting ofthe different raw material clusters provides a rangeof information on the occupation history of thesettlement.

From Lithics to Spatial and Social Organization 131

Promontory edge

N

0 3 m

Kettig“Im Rauental”

One equidistance

>365293–365220–292147–21974–1461–73

26/21

A

Variable equidistances

F–12E–56D–4C–20B–73A–35

All lithic artefacts

B

D

C

D

E

F

26/21

(a)

(b)

Figure 4. Kettig. Distribution of all lithic artefacts for (a) the entire site with an ‘‘equidistance’’ of N=73 and for (b) the artificial sub-areas(A–F) with variable ‘‘equidistances’’.

Plotting the artefacts by different size groups (e.g.<1 cm or 1–4 cm) merely mirrors the results for theoverall distributions and will not be presented here. In

contrast, plotting the different raw materials producesmore interesting results and allows an interpretation ofthe development of individual artefact accumulations.

132 M. Baales

A plot of the artefacts of Tertiary quartzite (Figure5(a)) shows that the densest concentration is at thesouthwest of the excavation (cf. the densest concen-tration of all artefacts, Figure 4(b), area E). Themajority of the Tertiary quartzite comprises very smallchips (<1 cm). In contrast with this situation, thepronounced concentration at the north of the siteshown by the plot of all materials (area A in Figure4(b)) disappears completely in Figure 5(a), sincethis concentration consists mainly of artefacts ofsiliceous slate (Figure 5(b)). Further artefacts of sil-iceous slate are found particularly to the northeast(area B) and, much less commonly, at the southwest(area E) of the site. Siliceous slate occurs only rarelyin the eastern test trench, where Tertiary quartzitedominates again.

Two clear concentrations of West European flint(Figure 5(c)) are present at the northeast of the site(area B). The most extensive of these overlaps with thatof the siliceous slate at this position. Two smallerconcentrations are visible to the south, one in the areaof the overall densest concentration of artefacts (areaE) and another to the west of this in area F. Themajority of finds in the latter concentration are small,normally fire-cracked chips. They probably representthe remains of only a limited number of artefacts whichwere burnt elsewhere and dumped at this locationwhere no other traces of fire were found.

A translucent type of flint can only be described aseither of Baltic or Tetange type. Both materials appearvery similar, although their respective source would bec. 100 km to the north in Saalian moraines or at alesser distance in the region west of Kettig. This flintshows a very specific distribution pattern (Figure 6(a)),with two clear scatters at the northeast (area B), welloutside the concentration of West European flint (seeFigure 5(c)).

The distribution of chalcedony artefacts (Figure6(c)) is also very distinct and reveals a dense accumu-lation again at the northeast of the site (area B),unfortunately partially disturbed by quarrying activi-ties before the discovery of the site. Further singleartefacts of chalcedony are scattered to the west andsouth, but the bulk of the chalcedony lies at the samelocation as the small northeastern concentration ofWest European flint (see Figure 5(c)) and peripheral tothe accumulation of Baltic/Tetange flint to the north(Figure 6(a)).

In summary, several, mostly distinct concentrationsof different lithic raw materials could be observed atKettig. Such dense scatters of debitage at Stone Agesettlements are generally attributed to two primarytypes of activities. Firstly, they can mark the actualsites of knapping activities, i.e. the locations at whichnodules were worked. A second possibility is that theyrepresent sites (‘‘dumps’’) at which material workedelsewhere was discarded or deposited.

Knapping locations are suggested as the explanationfor the concentrations of artefacts in the north and

northeastern parts of the site (areas A to D). Theselocations are mostly represented by single raw ma-terials (especially areas A, C, D and some spots in B)and, furthermore, are situated in area B close to thehearth(s) discussed below.

The very dense concentration of lithics in area E isclearly separated from the hearth(s) in area B. Follow-ing Behm (1983), the pronounced concentration ofthis material and the close spatial clustering speakclearly for an interpretation as a ‘‘secondary dump’’.Although a hypothetical knapping episode at thislocation cannot be ruled out (or recognized), thesubsequent use of area E as a ‘‘dump’’ where lithicwaste was discarded can be definitely recognized. Thisinterpretation is supported by subsequent, moredetailed spatial analyses of individual nodules of theraw materials.

The variable characteristics of the broad rawmaterial types allow the identification of specific units(‘‘nodules’’) which can be subjected to a further stageof analysis. Three examples are given here for Tertiaryquartzite. The main occurrence of a dense—in struc-ture and colour—homogeneous, grey quartzite (Figure7(a)) is in area B and the dense, southwestern concen-tration (area E). A few crested blades in area B identifythe knapping area of this material, although a core ofthis ‘‘nodule’’ was not found. The presence of thisdistinct Tertiary quartzite in the dense concentration ofarea E shows that some waste, particularly flakes andchips, produced at the knapping spot near a hearth atthe northeast of the site (area B) was subsequentlydumped several metres away to the southwest.

A yellow–grey Tertiary quartzite with large yellowinclusions and a coarse, off-white weathered cortexshows a totally different situation (Figure 7(b)). Thematerial is concentrated at the south of the site (areaD) and practically absent from the adjacent dump(area E). The presence of two cores allows the locationto be interpreted as a small knapping area at which lowintensity, primary flaking activities took place. Thisarea lies some distance away from the main, northeast-ern occupied area with the hearth(s) (area B) and it wastherefore not ‘‘cleaned up’’ or significantly disturbedby subsequent human activities.

A third situation is represented by a variety of brownquartzite with yellow flecks and water-rolled cortex.This material is concentrated at the very east of the site(Figure 7(c)) and the artefacts include a roughly pre-pared and still viable core with a refitting flake, whichwas nevertheless discarded. It is believed that theknapping of the brown quartzite was one of the lastactivities carried out at Kettig, and artefacts werefound clustered tightly together, rather than dispersed,since the site was deserted shortly afterwards. Only asingle scraper of this material was found in the easterntrench some distance from the concentration. Thisdistribution parallels that of the chalcedony (Figure6(b)), which is therefore also associated with thelast phase of site activity, an interpretation further

From Lithics to Spatial and Social Organization 133

N

0 3 m

Kettig“Im Rauental”

Tertiary quartzite

>190153–190115–15277–11439–761–38

26/21

26/21

26/21

Siliceous slate

>160129–16097–12865–9633–641–32

West European (“Meuse”) flint

>280225–280169–224113–16857–1121–56

(a)

(b)

(c)

Figure 5. Kettig. Distribution of all lithic artefacts of Tertiary quartzite, siliceous slate and West European (‘‘Meuse’’) flint.

134 M. Baales

supported by the concentrated accumulation of thismaterial.

N

0 3 m

Kettig“Im Rauental”

Flint of Baltic/Tétange type

>6049–6037–4825–3613–241–12

(a)

Chalcedony

>3025–3019–2413–187–121–6

(b)

Figure 6. Kettig. Distribution of all lithic artefacts of Baltic (Erratic)/Tetange type flint and chalcedony.

RefittingAlthough only a small amount of lithic material couldbe refitted, analysis of the conjoining lines providessome further interesting information on the settlementhistory and activities which took place at Kettig.

One hundred and twenty-two of the total of 24,098artefacts (3,834 artefacts >1 cm) were refitted, a pro-portion of only 0·5% (3·18%). They form 55 refittingcomplexes and provide a total of 67 conjoining lineswhich were plotted following the system of Cziesla

(1986, 1990). This gives a total of 40 sequential refits(Aufeinanderpassungen), 21 breakage refits (Anein-anderpassungen), five modification refits (Anpassungen)and one thermal refit (Einpassung eines Hitzeauss-prunges). The sequential refits thus dominate. (Thisapplies to all raw materials.)

The 65 plotted conjoining lines are mainly concen-trated in two areas (Figure 8). The majority (37 lines)lies to the northeast (area B), while 18 are found to thesouthwest, mainly linked to the dump (area E). Thetwo areas are connected by a c. 7·5 m long sequentialrefit of Tertiary quartzite, conjoining a small bladeletfragment at the southwest with a long and thick bladeat the northeast (Figure 8(a)). This is representativefor the transportation of material knapped at the

From Lithics to Spatial and Social Organization 135

N

0 3 m

Kettig“Im Rauental”

Homogeneous grey Tertiary quartzite

CbCbCb

Cb

Yellow–grey Tertiary quartzite

CtC

C

Brown Tertiary quartzite withwater-rolled cortex

C

SS

S

(a)

(b)

(c)

Figure 7. Kettig. Distribution of the larger artefacts of three raw material units (‘‘nodules’’) of Tertiary quartzite. C: core; Ct: core tablet;Cb: crested blade; S: end scraper.

136 M. Baales

northeast to the dump at the southwest of the site. Aconjoining line from the southwest dump extends some8·25 m to the northwest, an area in which only afew long conjoins were recognized, and is difficult tointerpret.

The northeast of the site (area B) is interpreted as themain area of site activity (as shown in more detail inthe following chapter), not least due to the types ofconjoins found here. While 20 of the 37 lines (54·05%)represent sequential refits (knapping), a further 17(45·95%) are due to breaks, which suggests that subse-quent human settlement activities at area B, especially‘‘trampling’’, also influenced the artefacts by breakinglithics lying on the surface.

The situation at the southwestern part of the site(area E) is different. The 12 (66·66%) sequential refitsrepresent the disposal of series of knapped waste at thedump. Only two of the 18 conjoining lines (11·11%) aredue to breaks, suggesting that, once deposited at thispart of the site, the artefacts were not subsequentlygreatly modified as a result of human activities such astrampling.

a

a 2

3

1

2

2

Figure 8. Kettig. Refitting lines of conjoining lithic artefacts. �, 14 m2 find; �—1, sequential refit; — · —2, breakage refit; · · � · ·3,

modification refit; – – � – –, thermal refit.

The Question of HearthsUnlike at the Federmessergruppen site of Niederbieber(Bolus, 1992), it was not possible to identify hearths atKettig as small areas of recognizably burned sediment.Nevertheless, it was already possible during excavationto observe two concentrations with numerous burnt

bone splinters and lithic artefacts at the northeast ofthe site (area B). The plot of white, calcined bonesplinters (representing the burning of bones at hightemperatures: Figure 9) shows that most of thismaterial lies at the northeast of the site, with its majordistribution in square 47/42. Most burnt fragments ofmammal teeth as fire-cracked artefacts lie in the samearea. Plotting the centres of distribution of the burntbone, tooth and artefacts, and also of thermally alteredporphyry, quartz and quartzite confirms the locationof a hearth at this part of the site (Figure 10).

A few metres to the southwest of this heart is asecond, less clearly defined accumulation of burntmaterial (Figure 10). Although many calcined bonefragments are present (Figure 9), they are here lessconcentrated than in the hearth and form a morediffuse satter. The same is true for burnt tooth frag-ments, while burnt artefacts were very uncommon.This situation might have two different explanations:

(1) It is possible that a fire burned here (not necessar-ily for long time) at the beginning of the occu-pation of the Kettig site, before the hearth insquare 47/42 was in use and the remains of theearlier hearth were dispersed and distorted bylater activities on the site. In this hypothesis, theunburned material found at this part of the site(especially lithic artefacts) would have been sub-sequently deposited when the older hearth was nolonger in use.

From Lithics to Spatial and Social Organization 137

N

0 3 m

Kettig“Im Rauental”

Calcined bone splinters

>510409–510307–408205–306103–2041–102

Figure 9. Kettig. Distribution of calcined bone splinters.

N

0 3 m

Kettig“Im Rauental”

Burnt lithic artefacts

Calcined bone splinters

Thermally altered quartzite cobbles, etc.

Thermally altered quartz and porphyry

Hearth (square 47/42)

Hearth/dump of burnt materials?

Figure 10. Kettig. Distribution of different burnt materials (lithic artefacts, calcined bone splinters, thermally altered cobbles of Devonianquartzite, quartz, porphyry, etc.).

(2) Alternatively, this area contains only materialintentionally cleared out (cleaning, maintenance)from the hearth in square 47/42, only a few metresto the northeast, which would have been in useduring the entire settlement process at Kettig.

It is difficult to decide which hypothesis is correct,but to the author, the first one appears more plausible.

Another category of burnt material is provided bynumerous fragments of quartz cobbles shattered bytheir repeated use as pot-boilers (Batchelor, 1979;Dittmann, 1990). Thermally altered quartz is particu-larly common in the same area as the majority of theburnt lithic artefacts. At the northeast of the site, theycan be interpreted as the waste from the final episodesof water-heating activity around the hearth in

138 M. Baales

square 47/42 (see Figure 10). Other, more restrictedconcentrations of burnt quartz fragments to the south(e.g. in area E, which is otherwise characterized as adump for lithic artefacts) can be interpreted as due tosimilarly discarding ‘‘exhausted’’ quartz fragments leftover from earlier episodes of boiling using the hearth(s)at the northeast of the site.

Interlude—Locating the Tent? The ‘‘Ring andSector’’ Method at KettigThe question whether the hearth identified in square47/42 was within or outside a dwelling structure can beinvestigated by application of the ‘‘ring and sector’’method developed by Stapert (1992). This analysis wasperformed earlier in Central Rhineland Federmesser-gruppen contexts at Niederbieber (Bolus, 1992; Stapert,1992), with negative results, and for two hearths atAndernach-Martinsberg (Stapert & Street, 1997),where, as at Kettig, no evident tent structures could berecognized. The analysis at Andernach showed thatone hearth was inside a tent of some 4 m in diameter.

At Kettig the ‘‘ring and sector’’ analysis of the coresand retouched artefacts (tools), with the centre of thehearth in square 47/42 and using distance increments of0·5 m, shows that it is not possible to recognize a ‘‘walleffect’’, showing that this hearth was not located insidea tent.

Furthermore, the analysis did demonstrate that vari-ous categories of tools behave in different ways. Endscrapers tend to be found closer to the hearth thanprojectile points, a feature already described forNiederbieber Area I by Stapert (1992: 83). As an expla-nation, Stapert suggests that, unlike the scrapers at thesomewhat older Upper Palaeolithic sites, their locationat Federmessergruppen sites does not reflect the spot oftheir use. Rather, the end scrapers were removedfrom their hafts (‘‘hafting and re-tooling’’) by softeningtheir adhesive in a fire and discarded in the ‘‘dropzone’’ close to the hearth. They would also be relativelyfrequently burned, a situation clearly substantiated atKettig around the hearth at square 47/42.

While the interpretation seems plausible for hearth47/42, the absence of adhesives on a sample of fivescrapers from Kettig investigated microscopically(although traces of adhesive were found on all fiveinvestigated projectile points; see Pawlik [in Baales, inpress; Baales, 1999]) and the overall diffuse distributionof scrapers, in more or less similar frequenciesacross the entire site, might speak against its generalapplicability.

Occupation Division at KettigOn the evidence of the flaked stone artefacts and, to alimited extent, on the basis of the other discussedmaterials, at least two phases can be identified duringthe occupation of the Kettig site:

(1) In the area of the hearth in square 47/42, certainraw materials used for stone tool production arefound in relatively discrete concentrations.Among these are a small but dense concentrationof Tertiary quartzite artefacts directly to thesoutheast of the hearth and the vast majority ofthe artefacts of Baltic/Tetange flint. These arte-facts are demonstrably the result of tool produc-tion around the hearth. Since the material is stillvery clearly concentrated (i.e. has not been dis-persed by a long subsequent period of humanactivity) and contains numerous fire-crackedspecimens, it is suggested that it represents ayounger phase of the occupation of the site whilethis hearth was in use.

Although concentrations of artefacts of chal-cedony and a particular type of brown Tertiaryquartzite somewhat to the southeast are locatedclose to the area of greatest human activityaround the hearth, they are not dispersed, and it isbelieved that these raw materials also belong tothe younger phase of settlement.

(2) Concentrations of artefacts of West Europeanflint and siliceous slate cover a larger, diffuse areato the south of the hearth in square 47/42 (Figure5(b)). They are believed to derive primarily froman older phase of settlement, possibly centredaround an associated hearth, and to have beendispersed by ‘‘trampling’’, etc. during the entiresucceeding phase of occupation.

The absence of artefacts of Baltic/Tetange flint,brown Tertiary quartzite and chalcedony (assigned tothe younger occupation phase) in area E suggests thatthe dump site reconstructed here should be associatedwith earlier settlement activities. Burnt quartz frag-ments found at the dump can most probably beattributed to early stone boiling events. The knap-ping areas of siliceous slate, and those of Tertiaryquartzite on the periphery of the site (e.g. Figure 7(b),(c)), probably also represent the older phase ofoccupation.

Summary of the Spatial Analysis and someGeneral Remarks on the Raw MaterialEconomyThis paper could only give a brief summary of certainaspects of the Federmessergruppen site of Kettig, thefull analysis of which will be presented in the finalpublication (Baales, in press). The main emphasis waslaid on the possibilities of using lithic artefact distri-butions for the recognition of internal site spatialorganization and fine chronological division of occu-pation phases. It was possible to differentiate severaldiscrete knapping areas of lithic artefacts, some ofwhich clearly relate to a hearth, and a ‘‘dump’’ wherenumerous lithic artefacts knapped elsewhere on the sitewere discarded.

From Lithics to Spatial and Social Organization 139

0

Per

cen

tage

pre

sen

t

10

20

30

up to 1

40

50

> 60–c.10020 40 60

X

71

4563

948

10 042

250881

245

337

1279

9112

Raw material resource distance (km)

Figure 11. Raw materials classed by their source. Kettig in compari-son with Andernach-Martinsberg (Federmessergruppen assemblageof the 1980s excavations). �, with Baltic/Tetange flint. Andernach(�), N=2·793; Kettig ( ), N=24·098.

Site history at Kettig can be divided into two succes-sive phases of occupation (which are, however, closelylinked), the younger of which is represented by series ofartefacts still located at the place where they wereknapped, while the older stage is demonstrated by rawmaterials found either more diffusely across the sitedue to their dispersal by subsequent human activity orin a secondary dump. Apart from these insights intothe spatial organization of a large Final Palaeolithicsettlement in the Central Rhineland, analysis of thedifferent lithic raw materials at Kettig allows someconclusions of a more general nature.

Final Palaeolithic Federmessergruppen sites in theCentral Rhineland show a striking use of exogenousraw materials (Figure 11) for manufacturing all kindsof tools, even though good quality raw material waspresent locally. ‘‘Exogenous’’, as used here, meansraw materials (cretaceous flint) obtained from sourcessome 80 km (and more) distant, especially thoselocated in the Meuse drainage area (Belgium andThe Netherlands) to the northwest, or from morainesof the Saalian glaciation to the north. The intro-duction of flint into the Central Rhineland can beobserved from the early Middle Palaeolithic onwardsand is particularly pronounced for the Upper Palaeo-lithic Magdalenian and for the Final PalaeolithicFedermessergruppen (Floss, 1994; Baales & Street,1996).

Seasonal information for both the Magdalenian andthe Federmessergruppen sites in the Central Rhinelandshows the availability of large game animals as a foodresource on a year-round basis. This implies thathunter–gatherers at both periods (separated by some2 ky) were potentially able to remain all year roundin the Neuwied Basin and its vicinity (Baales &Street, 1996, 1998). In this case, we need to ask whydid people change location and transport exogenousmaterials?

It is very probable that the reason for this mobilitywas the necessity of maintaining regular social contacts

with other hunter–gatherers in neighbouring regions.The need for such social interaction is made clear by arecent study of European population density duringdifferent pre- and protohistoric periods (Zimmermann,1996). Data presented by Vencl (1991), based onFedermessergruppen site density in Eastern Cen-tral Europe, suggest that the population of CentralEurope comprised some 0·02–0·03 individuals/km2

(Zimmermann, 1996: 58), only a tenth of a thousand ofthe recent value. This is, of course, only an estimate,but seems to be supported by data from sub-recenthunter–gatherer societies. Even without a detailedstudy, one can find in Pfizenmayer (1926: 248) a noticethat North Siberian reindeer hunters at the beginningof the 20th century showed a population density ofsome 0·04 individuals/km2 (close to the data discussedhere for the Federmessergruppen), although evensmaller values are known (cf. Weniger, 1982: 153,207).

In order to survive in such a sparsely inhabitedLate Glacial landscape, the small hunter–gatherergroups needed regular contact with other groups.These social networks allowed hunter–gatherers tointeract with other similar, local hunter–gathererunits, creating a form of higher-level ‘‘group iden-tity’’. Knowledge of and intermarriage between differ-ent local groups would in turn have provided aform of ‘‘insurance’’ in times of local subsistencecrisis by temporarily making available regionsoccupied by other (now ‘‘related’’) people, and alsohave made it possible to avoid inbreeding of the‘‘own’’ local group. Although established for socialreasons, once in existence, the networks alsoenabled the exchange and diffusion of knowledge ofnew technological innovations and thus incidentallyled to the development of archaeologically recog-nizable inter-regional technocomplexes (e.g. theFedermessergruppen).

In this model, lithic raw material curation andtransportation was simply a by-product of the move-ments initiated by highly necessary regular socialcontacts between the Central Rhineland Federmesser-gruppen groups with others in neighbouring regions ofNorthwestern Europe. All Federmessergruppen sitesin the Neuwied Basin have produced artefacts ofcretaceous flint from the Meuse catchment area tothe northwest. Although no Central Rhineland rawmaterials have yet been found on contemporary sites inBelgium or The Netherlands, a few Dutch sites of thepreceding Magdalenian have produced small numbersof Tertiary quartzite artefacts (often backed bladelets)presumed to have a Central Rhineland origin (Arts &Deeben, 1987: 55 f) and brought in by people from thesouth.

Whereas the Federmessergruppen contacts betweenthe Central Rhineland and the northwest have longbeen known, the recent recognition of Niederbieberof several nodules of Muschelkalkhornstein (Triassicchert; H. Lohr, pers. comm.) provides new evidence

140 M. Baales

for contact on a similar scale to the region aroundSaarbrucken, almost 200 km to the southwest.

AcknowledgementsMy thanks are specially due to the Deutsche Forsc-hungsgemeinschaft in Bonn-Bad Godesberg, whichmade possible the analysis of the Kettig site within theframework of a DFG Priority Programme Wandel derGeo-Biosphare wahrend der letzten 15.000 Jahre. A firstdraft of this paper, intended as a contribution to thisresearch programme, was presented in 1995 at the‘‘Analithic’’ meeting in Groningen, The Netherlands(Dr D. Stapert). Finally, many thanks to my colleagueDr M. Street, Neuwied-Monrepos, for discussion andimprovement of the text and for many fruitful past(and future) debates, and two unknown JAS refereesfor their useful, critical comments on the first draft ofthe paper.

ReferencesArts, N. & Deeben, J. (1987). On the northwestern border of

Late Magadalenian territory: ecology and archaeology of earlyLate Glacial band societies in northwestern Europe. In (J. M.Burdukiewicz & M. Kobusiewicz, Eds) Late Glacial Europe.Culture and Environment. Prace Komisji Archeologiczej 5.Wroclaw: Zaklad Narodowy im. Ossolinskich-Wydawnictwo,pp. 25–66.

Baales, M. (1994). Kettig (Kr. Mayen-Koblenz). Ein spatpalaoli-thischer Siedlungsplatz unter dem Bims des Laacher See-Vulkansim Neuwieder Becken. Ein Vorbericht. Archaologisches Korre-spondenzblatt 24, 241–254.

Baales, M. (1997). Kettig (Neuwied Basin, Central Rhine Valley,Germany): A Federmessergruppen site yielding organic artefacts.In (J.-P. Fagnart & A. Thevenin, Eds) Le tardiglaciaire en Europedu Nord-Ouest. Actes du 119e Congres national des societeshistoriques et scientifiques, Amiens 1994. Paris: CTHS, pp. 579–588.

Baales, M. (1998). Zur Fortfuhrung der Ausgrabungen aufdem spatpalaolithischen Siedlungsareal von Niederbieber (StadtNeuwied). Ein Arbeitsbericht. Archaologisches Korrespondenzblatt28, 339–356.

Baales, M. (1999). Neue Untersuchungen zum Spatpalaolithikumdes Neuwieder Beckens: Einige Aspekte des Federmesser-Fundsplatzes Kettig, Kr. Mayen-Koblenz. In (E. Cziesla, T.Kersting & S. Pratsch, Eds) Den Bogen spannen . . . Festschriftfur Bernhard Gramsch zum 65. Geburtstag. Beitrage zur Ur- undFruhgeschichte Mitteleuropas 20. Weissbach: Beier & Beran,pp. 55–66.

Baales, M. (in press). Kettig (Kr. Mayen-Koblenz)—Ein Siedlung-splatz der Allerød-Zeit im Neuwieder Becken. Mainz: Romisch-Germanisches Zentralmuseum Mainz.

Baales, M. & Street, M. (1996). Hunter–gatherer behaviour in achanging late glacial landscape: Allerød archaeology in the CentralRhineland, Germany. Journal of Anthropological Research 52,281–316.

Baales, M. & Street, M. (1998). Late Palaeolithic backed-pointassemblages in the northern Rheineland: current research andchanging views. Notae Praehistoricae 18, 77–92.

Baales, M., Bittmann, F. & Kromer, B. (1998a). Verkohlte Baumeim Trass der Laacher See-Tephra bei Kruft (Neuwieder Becken).Ein Beitrag zur Datierung des Laacher See-Ereignisses und

zur Vegetation der Allerød-Zeit am Mittelrhein. ArchaologischesKorrespondenzblatt 28, 191–204.

Baales, M., Mewis, S. U. & Street, M. (1998b). Der Federmesser-Fundplatz Urbar bei Koblenz (Kreis Mayen-Koblenz). Jahrbuchdes Romisch-Germanischen Zentralmuseums Mainz 43 (1996),241–279.

Batchelor, D. (1979). The use of quartz and quartzite as cookingstones. In (G. Bosinski, Ed.) Die Ausgrabungen in Gonnersdorf1968–1976 und die Siedlungsbefunde der Grabung 1968. DerMagdalenien-Fundplatz Gonnersdorf 3. Wiesbaden: FranzSteiner, pp. 154–165.

Behm, J. A. (1983). Flake concentrations: distinguishing betweenflint working activity areas and secondary deposits. Lithic tech-nology 12, 9–16.

Bogaard, P. van den & Schmincke, H.-U. (1985). Laacher SeeTephra—a widespread isochronous late quarternary tephra layerin Central and Northern Europe. Geological Society of America,Bulletin 96, 1554–1571.

Bolus, M. (1992). Die Siedlungsbefunde des spateiszeitlichen Fund-platzes Niederbieber (Stadt Neuwied). Monographien desRomisch-Germanischen Zentralmuseums 22. Mainz & Bonn:Dr. Habelt.

Bosinski, G., Street, M. & Baales, M. (Eds) (1995). The Palaeo-lithic and Mesolithic of the Rhineland. In (W. Schirmer,Ed.) Quaternary Field Trips in Central Europe 15, Vol. 2. 14thINQUA-congress. Berlin, Munich: Dfr Friedrich Pfeil, pp. 829–999.

Cziesla, E. (1986). U} ber das Zusammenpassen geschlagener Steinar-tefakte. Archaologisches Korrespondenzblatt 16, 251–265.

Czeisla, E. (1990). Siedlungsdynamik auf steinzeitlichen Fundplatzen.Methodische Aspekte zur Analyse latenter Strukturen. Studies inModern Archaeology 2. Bonn: Holos.

Dittmann, A. (1990). Das Kochen mit Steinen. Sonderschriften desFroebenius Instituts 9. Frankfurt/Main & Wiesbaden: Froebenius.

Floss, H. (1992). Bonn-Muffendorf und die Chalcedone des Rhein-landes. Naturliche Vorkommen und palaolithische Nutzung.Jahrbuch des Romisch-Germanischen Zentralmuseums 36 (1989),159–177.

Floss, H. (1994). Rohmaterialversorgung im Palaolithikum desMittelrheingebietes. Monographien des Romisch-GermanischenZentralmuseums 21. Mainz & Bonn: Dr Habelt.

Ikinger, A. (1996). Bodentypen unter Laacher See-Tephra imMittelrheinischen Becken und ihre Deutung. Mainzer geowissen-schaftliche Mitteilungen 25, 223–284.

Joris, O. & Weninger, B. (2000). Radiocarbon-calibration and abso-lute chronology of the Late Glacial. In (B. Valentin, P. Bodu & M.Christensen, Eds) L’Europe centrale et septentrionale au Tardi-glaciaire. Confrontation es modeles regionaux de peuplement.Colloquium Nemours 1997 Memoires du Musee de prehistoired’Ile-de-France 7. Nemours: A.P.R.A.I.F., pp. 19–54.

Kalthoff, D. C. (1998). Die Kleinsauger (Mammalia) der FundstelleKettig (Rheinland-Pfalz, Deutschland) im Rahmen derallerødzeitlichen Saugetierfauna Mittel- und Suddeutschlands.Palaontologische Zeitschrift 72, 407–424.

Pfizenmayer, E. W. (1926). Mammautleichen und Urwaldmenschen inNordost-Sibirien. Leipzig: F. A. Brockhaus.

Schwabedissen, H. (1954). Die Federmesser-Gruppen des nordwesteu-ropaischen Flachlandes. Zur Ausbreitung des Spat-Magdalenien.Offa-Bucher 9. Neumunster: Wachholtz.

Stapert, D. (1992). Rings and sectors: intrasite spatial analysis ofStone Age Sites. Ph.D. Thesis. University of Groningen.

Stapert, D. & Street, M. (1997). High resolution or optimumresolution? Spatial analysis of the Federmesser site at Andernach,Germany. World Archaeology 29, 172–194.

Street, M. & Baales, M. (1997). Les groupes a Federmesser del’Allerød en Rhenanie centrale (Allemagne). Bulletin de la SocietePrehistorique Francaise 94, 373–386.

Street, M., Baales, M. & Weninger, B. (1994). Absolute Chrono-logie des spaten Palaolithikums und Fruhmesolithikums imnordlichen Rheinland. Archaologisches Korrespondenzblatt 24,1–28.

From Lithics to Spatial and Social Organization 141

Veil, S. (1982). Der spateiszeitliche Fundplatz Andernach-Martinsberg. Germania 60, 391–424.

Vencl, S. (1991). On the importance of spatio-temporal differences inthe intensity of Palaeolithic and Mesolithic settlement in CentralEurope. Antiquity 65, 308–317.

Weniger, G.-C. (1982). Wildbeuter und ihre Umwelt. Ein Beitrag zumMagdalenien Sudwestdeutschlands aus okologischer und ethno-

archaologischer Sicht. Archaeologica Venatoria 5. Tubingen:Archaeologica Venatoria.

Zimmermann, A. (1996). Zur Bevolkerungsdichte in der Urgef-schichte Mitteleuropas. In (I. Campen, J. Hahn & M. Uerpmann,Eds) Spuren der Jagd—Die Jagd nach Spuren. Festschrift H.Muller-Beck. Tubinger Monographien zur Urgeschichte 11.Tubingen: Mo Vince, pp. 49–61.