MESOLITHIC HEARTH-PITS IN THE VEENKOLONIEN (PROV. GRONINGEN, THE NETHER­ LANDS), DEFINING A SPECIFIC USE OF FIRE IN THE MESOLITHIC

H.A. Groenendijk Bi% gisch-Archae% gisch ll/stitllllt, Gral/il/gen, Nether/al/ds

ABSTRACT: The principles of construction and use of Mesolithic hearth-pits are examined in the Groningen part of the Veenkolonien, a peat reclamation district where a uniform 'Mesolithic' landscape came to light when the peat beds were dug away. Put in a broader context, these Mesolithic hearth-pits seem to have been 'designed' for food-processing mainly and used in specific areas within the settlements. They are tentative ly compared with Late Palaeolithic hearths from adjacent areas, in an attempt to demonstrate that a new type of hearth emerged at the beginning of the Mesolithic. Post-glacial environmental adaptation may be reflected in these hearth-pits according to the 'form follows function' principle.

KEYWORDS: Late Palaeolithic hearths, Mesolithic headh-pits, Younger Coversand II landscape, flint distribution, non-domestic hearths, hearth-pit experiment.

1. INTRODUCTION lithic and Mesolithic, we are confronted with a lacuna in Ouf knowledge. Only traces af the use af Our knowledge af the Late Palaeolithic and Mesoli­ fire may serve as 'index fossils'. thic in the Netherlands is based mostly on finds Hearths, more than anything else, are clear from dry, inorganic, Pleistocene and Early Holo­ indications af the use af fire. 'Mesolithic hearth­ cene deposits. The camp sites first discovered all lie pits' especiaIly are remarkable for their homogenei­ in coversand formations, as the search strategy used ty in shape and contents. Their'carbonized botani­ to focus exclusively an these easily accessibie areas. cal contents are particularly suitable for analyses. Above all, elevations in the terrain were investi­ They may yield radiocarbon dates with compara­ gated. The state af Late Palaeolithic and Mesolithic tively narrow standard deviations. Investigation af research in fact still bears the marks af this bias the hearths' shape and contents may provide in­ towards surface collections. This selective sampling formation about their use, while identification of is ane of the factors responsibie for the fact that in the kinds af wood may contribute to a reconstruc­ the Netherlands no camp sites are known from tion af the Mesolithic vegetation. organic deposits, while in Denmark, an the con­ An analysis of the shape and contents of this type trary, the emphasis in research has for a great part af Mesolithic hearth together with those af Palaeo­ been on Maglernosian sites in low-lying areas. lithic hearths can be a method af comparing aspects The discrepancy between the numbers af known af the use af fire in the two periods and more Late Palaeolithic and Mesolithic sites in coversand particularly of investigating the Late Palaeolithic areas is due to the same tendency. Mesolithic sites in and the Mesolithic partition and use af space in coversand formations have a far greater chance af settlements. lying at the modem surface than do their (mostly) The data are inevitably distorted by the small dune-covered Late Palaeolithic counterparts. Sur­ number af Late Palaeolithic hearths uncovered in face sites af COUfse are the most frequently ex­ the Northwest-European Plain, against the pro­ cavated. Sites of this categary do, however, have a fusion af Mesolithic hearth-pits. Moreover it is major drawback: they seldom represent a single unknown what kinds af hearth are being compared; occupation phase. At such sites Late Palaeolithic an analysis af the two groups may at best bring to elements may lie intermingled with Mesolithic as­ light functional differences. semblages and aften younger anes as well. Dis­ The findings from the many hearth-pits that have tinguishing the various constituents is very difficult come to light in the district known as the Veen­ with a large part af the flint material. However, kolonien ('Peat Calanies'), in the east af the pro­ ground features may contribute towards a positive vince af Groningen, in this study provide the identification af the various traditions represented database for the 'Mesolithic approach'.! The in­ at the site. Yet when it comes to the Late Palaeo- vestigated area (fig. l) is characterized by a sub-

85 86 H.A. GROENENDIJK

Fig. I. Location ofthe study area (outlined) in the provinee of Groningen. The location ofthree frequently mentioned sites is also given (l

= NP3; 2 = SSI; 3 = S6).

stratum of coversands with a relief that is typical of It is assumed that the coversand landscape as it coversand; locally it is a confused pattern of ridges emerged during the reclamation of the peatlands and depressions. Yet the differences in height never represents the Mesolithic surface. During and im­ exceed two metres. As a result of poor drainage, this mediately after the Mesolithic occupation, some coversand region became' boggy and inaccessible to local erosion will have taken place through the Mesolithic people roughly from the Early Atlantic action of water and perhaps some deflation and period onward. Eventually it- was covered by a sedimentation through the action of wind. blanket of peat many metres thick. Now that the Flint artefacts have so far shown up both in the peat has been dug aWay, the Late-Glacial Younger A2 and B2 horizons of the podzol profile, !lnd just Coversand II lies at the surface almost everywhere. occasionally in the B3 horizon. No buried palaeo­ The sites that so far have been radiocarbon-dated sols were found. Sites with erosion gullies as well as 3 point to an occupation phase lasting from c. 9400 to sites with an intact old surface were encountered. 7500 BP. As ilO younger dates have been recorded, A few sites do contain some elements reminiscent this may have been followed by the widespread of the Tjonger (Federmesser) industry (Molerna, 2 development of bogs. The flint assemblages too 1988), bt-lt thc distribution of such sites is limited to suggest that this coversand region was visited no a small area. Their litho-stratigraphy has not yet more after the Early Atlantic. Nothing is known been investigated. about the possibIe exploitation of the area by Thus it can be said that in the major part of the Mesolithic man during the growth of the peat Veenkolonien the Mesolithic belongs in the Pre- blanket. Mesolithic hearth-pits in the Veenkolonien 87 boreal and Boreal, and that younger contamination c. 1 m and the other c. 0.5 m in diameter, lay one can be ruled out.4 Not only the unmixed character meter apart, but yet were connected through an area but also the early date of the Mesolithic sites in thes� of sand coloured dark by its admixture of charcoal. parts were incentives to investigate the Mesolithic This layer of dark sand nowhere was thicker than hearth-pits. 0.10 m. Its contours suggest that both hearths had peripheral zones of charcoal of c. 1.5 and 1 m across respectively, which partially overlapped. Outside 2. THE NATURE OF THE HEARTHS these zones the soil contained no finds. The dark

sand may have been the result of clearing-out ' hence 7 2.1. Late Palaeolithic hearths repeated use, of the hearths. A third hearth also featuring stones with charcoal and flint art�facts From the Netherlands two Hamburgian hearths among them, lay at a distance of c. 9 m from the and a single Ahrensburgian example are known. . other two. The documentation allows little more Even If we add tue other Late Palaeolithic hearths of than a surmise that this possibie hearth was les s t?e Northwest-European "Plain that have been pub­ than 1 m across. No radiocarbon dates are avail­ I shed so far, the total number still is not signi­ � abie. flcantly greater. In this section, the published At Rissen, near Hamburg, a hearth was un­ hearths from comparable lowland contexts (the covered amidst an Ahrensburgian assemblage. On­ Netherlands, northwestern Germany) will be briefly ly a plan of the hearth has been published (Rissen discussed. 14a; Schwabedissen, 1954). This shows a crescent­ A well-preserved and carefully excavated ex­ shaped patch of charcoal, with a maximum diam­ ample is the Hamburgian hearth of Oldeholtwolde eter of c. 2 m. A description was given of an province of Friesland (Stapert, 1982; Stapert et al.: Ahrensburgian layer with a thickness ofO.05 to 0.10 1986). The charcoal contained in the hearth was m, which contained the hearths (apparently there ra�iocarbon-dated to 1l,540±270 BP (GrN-I0274). w�re several, although only one was illustrated). Thls heart� lay at the centre of a Hamburgian flint Fhnt artefacts were found both within and outside concentratIOn. The hearth itself consisted of a the immediate vicinity of the described hearth. The heartshaped arrangement of flat stones, c. 1.5 m charcoal, being Pinus, actually suggests a date across, at the centre of which was a hollow mea­ younger than Ahrensburgian; no radiocarbon date suring 0.35 by 0.50 m and with a depth of 0.10 m. is available (pers. comm. J.N. Lanting). Traces of the action of fire on some of the hearth The Late Palaeolithic hearths mentioned above stones indicated that the hearth had been cleared sealed between coversand deposits, may all b� out and re-used several times. r garded as closed finds. Therefore they are par­ A possibie second hearth of the same period was � tlcularly useful for morphological comparison. found at the Hamburgian site of Luttenberg pro­ Table 1 summarizes some relevant characteristics. vi ce f Overijssel (Stapert, 1986; Stapert '& de � � Hamburgian hearths predominate in this cata­ Vnes, In prep.) There are indications that stones logue. :Vhat characterizes this fairly homogeneous were a constructional part of the hearth. Because of group IS the use of stones, and usually a wide a la ter disturbance it was not possibie to reconstruct diameter combined with a shallow depth. The two the exact size of this hearth. Ahrensburgian hearths of Geldrop and Rissen In Geldrop, province of Noord-Brabant an c ntained no stones, and morphologically they Ahrensburgian hearth was encountered in a se�tion � dlffer from each other. Geldrop I is the only hearth (Geldrop I; Wouters, 1957; 1983). There were no that appears to have been cut into the ground. Some stones, ?ut in the somewhat wedge-shaped 'hearth­ . , of the hearths demonstrably were the focus of flint­ PIt , whlch was c. 0.5 m deep and c. 0.8 m wide at the working activities (Oldeholtwolde, Querenstede 1 top, not only charcoal was found'but also calcined and 2, 3?); others at any rate lay within con­ animal bone;· at its edge lay some red ochre.5 The centrations of mnt artefacts (Luttenberg, Rissen hearth lay within a charcoal-rich zone from which 14a). Late Palaeolithic hearths genera Ily are fea­ it was not ve ry distinct. Radiocarbon a�alysis of the tures af ha�itation sites ('domestic hearths'); those charcoal produced a date of 10,960±85 BP (GrN- . mentIOned In table 1 all are associated with flints. 1059).6 However, this must be qualified in the sense that Three hearths Iying close together were un­ often Qnly the hearth and its immediate surround­ covered near Querenstede in Northwest Germany ings were investigated. It is for this reason that we (Zoller, 1963). Given the associated flint finds they cannot say whether in these parts Late Palaeolithic must belong to the Hamburgian tradition. All three hearths for other purposes may not have been of them were marked by a configuration of stones constructed at the periphery of the habitations to among which lay charcoal- most of it at the centre � which they yet belonged} and flint artefacts. Two of the hearths, one of them 88 H.A. GROENENDIJK

Table I. Morphology and associated finds of Late Palaeolithic hearths from the Northwest-European Plain.

Site Tradition Structure Dimensions in m Contents

Oldeholtw(Jlde (NL) Hamburgian Stone lining of Diam. I.5, Charcoal (Salix), of central pit depth 0.1; bumt + with peripheral central pit unbumt flint stones 0.350.50.1

Luttenberg (NL) Hamburgian Stones ? Charcoal, bumt flint

Geldrop I (NL) Ahrensburgian Pit 0.8 wide Charcoal, 0.5 deep bumt fl int, calcined animal bone, red ochre

Querenstede Hamburgian Stone concentration Diam. c. I, Charcoal, (BRD), hearth I in shallow depression 0.1 deep 'flint'

Id., hearth 2 Hamburgian Id. Diam. c. 0.5, Id. 0.1 deep

Id., hearth? 3 Hamburgian Stone concentration Diam. under I Id.

Rissen 14a Ahrensburgian Shallow depression? Diam. c. 2 Charcoal (Pi nus), (BRD) 'flint'

2.2. Mesolithic hearth-pits would be used to support its supposed date. Their identificatiol) as hearth-pits is widely accepted, as is Focussing on 'Mesolithic hearth-pits' in order to their Mesolithic origin, although arguments to compare them with Late Palaeolithic hearths we are connect them with flint industries on the same site confronted with heavy restrictions in several re­ often are lacking. spects. Not only is the abundance of excavated In the folIowing paragraphs an attempt is made hearth-pits from the Mesolithic all over the Pleisto­ to list some aspects of Mesolithic hearth-pits found cen e coversands of the Northwest-European Plain in the Veenkolonien, to arrive at a more specific in sharp contrast with the number of known Late iqea of how fire was used in this type of hearth. Palaeolithic hearths, but the knowledge about any Since the observations started in 1982, a great other types of hearth one might expect on Meso­ number of 'classical' Mesolithic hearth-pits have lithic sites is ve ry poor. The hearth-pits most been uncovered in the Veenkolonien; over a hun­ probably formed only part of the Mesolithic fea­ dred of them have been analysed.'o tures relating to the use of fire.9 One advantage of focussing on the hearth-pits is that they form a large 2.2.1. Frequency of occurrence and location and homogeneous group. Although being a very common phenomenon Theoretically, the chance of discovery of Mesolithic and occurring in many excavations (sometimes even occupation in the Veenkolonien is much greater considered as unwelcome archaeological by-pro­ than with part of the Late Palaeolithic, as the ducts as they tend to contaminate radiocarbon predominantly Younger Coversand II surfaces did samples from other periods), Mesolithic hearth-pits not stabilize until the beginning of the Mesolithic. have rarely been the subject of scientific study as to In the many man-made exposures in the Veen� form and function. Their primary analyticai func­ kolonien only habitation at the surface of the tion has been to provide reliable radiocarbon dates y ounger Coversand II has been observed. The few for the associated Mesolithic flint industries and sites with Late Palaeolithic elements mixed in with thus they have in fact provided the absolute chrono­ the Mesolithic flint assemblages also are known to logical framework for the Mesolithic stages in the have been surface collected. At several Mesolithic Netherlands. sites where excavations took place, hearth-pits were When occurring amidst Late Palaeolithic flint encountered. Table 2 shows that their occunence assemblages their dates were considered too young seems to be restricted to the top and slopes of the to be of a Late Palaeolithic origin and thus they largest sand dunes. Smaller and lower dunes car­ would be called Mesolithic hearths; when ac­ rying Mesolithic flint concentrations did not con­ companied by a Mesolithic flint industry they tain hearth-pits. Mesolithic hearth-pits in the Veenkolonien 89

Table 2. Frequency and location of hearth-pits on Mesolithic sites in the Veen kolonien.

Site/Code M' excavated Number of hearth-pits Geomorphology of the part uncovered

NP3 2500 38 Top/upper slope of elongated, pronounced dune

NP4 250 Top/upper slope of elongated, pronounced dune

Vm24 520 Lower slope of elongated, pronounced dune

HSI6 170 Edge/upper slope of extended elevation

HSI7 160 6 Edge/upper slope of extended elevation

HS22 390 7 Upper slope of elongated, pronounced dune S6 485 28 Top/upper slope of pronounced dune

S51 750 40 Top/upper slope of elongated, pronounced dune

S60 150 Top of low, small dune

S67 200 Low, small dune

Within a time-span of less than 2000 years of the artefacts in relation to the distribution pattern of Mesolithic according to the radiocarbon dates hearth-pits and on the hearth level there is the available at present, the digging of hearth-pits at analysis of their contents. Results of the latter are some sites proves to have been a frequent activity. discussed in section 2.2.5; the approach on site level The larger and fairly pronounced elevations in the is discussed in this section. coversand landscape of the Veenkolonien seem to If one wants to demonstrate non-coincidence of be favourable locations for the settlements con­ the distribution offlint artefacts and hearth-pits this taining hearths under discussion. Mesolithic flint is only possibie if both categories exclude each other concentrations also occur on the smaller elevations, spatiaIly in two dimensions, both horizontal and although there no hearth-pits were observed. Diffe­ vertical, or if the radiocarbon dates of the hearth­ rent locational requirements for hearth-pits com­ pits differ clearly from the accepted age of the pared to flint-working areas may be the cause of 'associated' flint industry." this, though the argument cannot be further sub­ From the Veenkolonien an example of non­ stantiated. coincidence of the horizontal distribution of flint artefacts and hearth-pits is given here (apart from a few stones, flint is the only artefact category ob­ 2.2.2. Non-coincidence with flint industry? served in this area). At the extensive Mesolithic site In their introduction to the Mesolithic Lanting and NP3, part of which was excavated, four or five main Mook (1977) warn that the dates for Mesolithic flint concentrations occurred. In the SW part of the hearth-pits are frequently not in accordance with site lies a small concentration with over 100 pie ces the typological assignments of the flint tools from per m2; at the centre and SW of it are two partly the immediate surroundings. Repeated occupation overlapping concentrations, with maxima of 49 and 2 of a site is mentioned as a main reason, contamina­ 24 pieces per m ; in the NE lies a minor con­ 2 tion of the charcoal being of little influence (J.N. centration of up to 15 pieces per m ; and the far NE Lanting, pers. comm.). With this point of view I has a probably limited concentration with a maxi­ 2 totally agree; exactly because the charcoal, lying at mum of 68 pieces per m • In the excavation the the bottom of these deep hearth-pits, will quite soon limits of flint occurrence were reached all along the have been covered in, it is unlikely to be con­ steep slope of the coversand dune in the SE, as well taminated. as in the NE, N and NW on the top of the dune (fig. There are two levels of observation: on the site 2). For the uncovered part of the site the distribu­ level there is the distribution of Mesolithic flint tion of hearth-pits does not correspond with these 90 H.A. GROENENDIJK

NP3 1984

5m , OJ NOT EXCAVATEO

Iill • • • 8090'30,. • • m 8230±45� •

OJ [I] m ill II] [I] OJ [I] [I] III

@] II] [IJ @]

D [I] ill [I] II] UD UD II] ill D D

UD

D m @] Iill !ill !Il !Il D

�

NOT llli EXCAVATEO

• • • • � � [j]J D NOT • llil • • • • • • llli

• [@ • •

EXCAVATED • "9110:':45 •

!Il tB •

@] m @] III m !Il

Fig. 2. Excavation plan ofthe Mesolithic site NP3, gemeellIeNieuwe Pekela. The extent afthe site was tested by pits of ane square metre; the area eventually uncovered is heavily outlined. af the ground features only the Mesolithic hearth-pits are shawn. Legend: I. test pits with the total number afOint artefacts (blank squares: no artefacts; black squares: more than 100 artefacts; 2. Mesolithic hearth-plts, three af which are radiocarbon dated (dates BP); 3. nest of c. 25 pieces af flint occurring together ('coaking stones'). Mesolithic hearth-pits in tlle Veenkoloniiin 91

Fig. 3. Part of the uncovered area of the Mesolithic site S51, gemeellIe Stadskanaal. The black spots are hearth-pits measured exactly; open circles: hearth-pits whose location could not be measured exactly, but whose measuring error does not ex­ ceed one metre. In the southeastem part the observed cIuster of nine hearth­ pits. Underlined figures are radiocar­ bon dates; numbers in italic beside the hearth-pits indicate the total number of bumt flint artefacts in their con­ tents.

flint concentrations, although the overall distribu­ tion of flint artefacts largely covers the distribution 8090'30• • of the hearth-pits. Yet several hearth-pits lie in an . : 8230'�S , • area where no flint was found. The picture of site @ NP3 appears to indicate a multiple componency ej Sm with a considerable spread in time, which is cor­ roborated by the radiocarbon dates from hearth­ NP3 pits both from a group of four in the N part and 1984 from a hearth-pit in the S part. This spread in time of hearth-pits at any rate, occurs at other Mesolithic Fig. 4. Detail of the excavation plan of site NP3 (see fig. 2) with sites in the Veenkolonien too. the cIuster of four hearth-pits. Underlined figures are radio­ carbon dates; the number in italic indicates that this hearth-pit With excavations lacking detail similar to NP3, it contained three bumt flint artefacts, while the others were sterile. is impossible to say whether hearth-pits together with low concentrations of flint do or do not form behavioural units. The discussed examples can only demonstrate the possibility of single (groups of) hearth-pits occurring outside flint concentrations. Furthermore, an analysis of the contents of It is this which seems to be in contrast with the few Mesolithic hearth-pits demonstrates that some of Late Palaeolithic hearths observed; they all feature these are nonetheless associated with contempora­ a flint concentration or even form the focus of flint neous flint working (see section 2.2.5). working (see 2.1). This seemingly different position Finally , it is very likely that the chosen examples of the hearth in relation to flint concentrations in from the Late Palaeolithic and the Mesolithic the Mesolithic and in the Late Palaeolithic is an represent two different types of hearths; testing important observation. However, some factors that their position in relation to flint concentrations is may have affected this picture must be taken into affected by their numerical unequality and the fact account. There are the different find circumstances, that excavations of Late Palaeolithic hearths mostly as for the Mesolithic a difference in the vertical cover only a very limited area. position of flints and hearth-pits, which are asso­ cia ted with and have their origin at the same 2.2.3. C1usters of hearth-pits surface, can be noticed: of Mesolithic hearth-pits generally only the base is observed, at a level where Considering the density of hearth-pits on some flint may be absent. Fortunately in the Veen­ Mesolithic sites, it is striking that in spite of the long kolonien the reconstruction of the old surface has time-spa n of intermittent habitation hardly any been possibie in some cases (see section 2.2.4). hearth transects another. It looks as if abandoned. 92 H.A. GROENENDIJK hearth-pits could be recogmzeCl for quite a while simultaneously. Two radiocarbon dates from and that re-use of the same pit was not the practice. hearths lying less than 4 m apart (6750±75 BP and an the other hand, the way som e hearth-pits c1uster 6870±85 BP) support the suggested con tempo­ is suggestive of gro ups of contemporaneous hearths raneity of the group. an one side the edge of the on Mesolithic camp sites. We observed this on the concentration was reached. Clearly dug within a sites S51 (fig. 3) and NP3 (fig. 4). narrow time-span, the delimited group of hearth­ an S51 nine hearth-pits were found in a c10se pits is referred to by the excavator as a 'cooking group, lying apart from the others, in an area of c. centre', although this is supported only by ethno­ 3.5 m diameter. None of the hearths transected graphic parallels. another. Some were faintly outlined, others showed clear contours (most probably depending on the Thus Mesolithic hearth-pits may occur in con­ amount of charcoal left after burning, see below; temporaneous c1usters of varying dimensions; their formerly erroneously interpreted as a difference in possibie relationship with the distribution of iso­ age). The largest hearth-pits were also richest in lated hearth-pits at the same site was not in­ charcoal. Two of these have been radiocarbon vestigated here. The intensity of the discoloration dated to check their contemporaneity; the results caused by charcoal in the fill of the pit apparently are 7615±40 BP (GrN-13747) and 7480±40 BP does not allow any conclusion with regard to age. (GrN-13748). The calibration curve in the reach of these dates shows a steep slope (Kromer et al., 2.2.4. Similarity of hearth-pits 1986), so that after correction the contemporaneity of the two hearths remains credible. This together Two aspects of Mesolithic hearth-pits determine the with the similarity of contents and the similarity of similar appearance of the hearths involved, their the rest of the hearths within the c1uster (e.g. the low apparent great number and their good conserva­ quantities of burnt flint artefacts found in them) tion: (a) their similar construction, especiaIly the indicates a clearly delimited area where fire was narrow range of variation in depth in relation to the used. The burning of fires need not have taken place old surface, and (b) the appearance that these in one session, but in any case the place must have hearth-pits were not cleared out. been recognizable when people· came to reoccupy a) Similar construetion. Recognition of Mesolithic it. 12 hearth-pits depends heavily on the stratum exposed The same may apply to an area on NP3 with four in relation to the Mesolithic surface. In the Veen­ equally large hearth-pits, one of them faintly out­ kolonien the greater part of the hearth-pits were lined, lying together in the form of a rectangular only recognized in the B3 horizon of the podzol trapezium. The spot measures c. 2.5 m across. Two profile, c. 0.40-0.50 m beneath the top of the hearths were radiocarbon dated: 8230±45 BP (GrN- coversand in which this profile developed. This is 13750) and 8090±30 BP (GrN-15313), which for the because genera Ily only the dark, charcoal-flecked same reason as the dates from S51 suggest con­ base of a hearth-pit is noticed, whereas the upper temporaneity of this group of four. part of the pit remains invisible by lack of charcoal Similar observations were made at Mesolithic and as a result of soil development.13 In the sites outside the study area. A.D. Verlinde ex­ Veen kolonien we encountered several hearth-pits in cavated 'numerous' hearth-pits adjacent to possibly section (fig. 5), allowing us to reconstruct their Mesolithic grave-pits at Marienberg, province of original size and depth. An average depth of 0.40- Overijssel (Verlinde, 1982; van Es et al., 1988). 0.50 m for the Mesolithic hearth-pits, as had been Neither the hearth-pits, nor the six grave-pits estimated aiready from their predominant position showed transections, in spite of the density of the within the B3 horizon was thus confirmed (the features observed. Five hearth-pits were radio­ geological arguments for believing that in the carbon dated; they cover a time-span of 6195±35 Veenkolonien the top surface of the coversand BP to 6290±45 BP (uncalibrated), which together dunes will generally represent the Mesolithic sur­ with the archaeological evidence suggests a con­ face, have been discussed in the Introduction). The temporaneous burning of fires. deep-Iying fires must have been less liable to In S tikken, West Germany, l.l. Assendorp drew contamination with other organic material such as attention to a ve ry dense c1uster of hearth-pits in a charcoal from other hearths, than would shallow limited area of c. 220 m2 underneath a barrow on a pits and depressions. coversand elevation (Assendorp, 1985). Over 70 The homogeneous morphology of the hearth-pits hearth-pits were observed; they showed hardly any in the Veenkolonien (see fig. 5) sets them apart from transection. Furthermore the excavation plan sug­ other ground features. The bowl-shaped, but more gested that several among them lay in pairs, as if often lenticular base forms the bottom of a shaft derelict hearths could easily be recognized but re­ that varies in diameter from 0.40-0.80 m. In fact the use wao avoi,ded, or that these hearths burned hearths generally measure 0.40-0.60 m across and Mesolithic hearth-pits in the Veenkolonien 93

a

b

c

d

Fig. 5. Hearth-pits with an 'intact base' as they were observed in section (a, b: site NP3; c, d, e: site S51). Legend: hatehed are the areas where the main horizons of the humus pod­ zol were recognized; l = A2; 2 = 82, 3 = 83. O,5m

only exceptionally reach a width of 0.80 m. The open fire is finished, the charcoal may continue to diameters of Mesolithic hearth-pits outside the glow for hours. ControIling the fire is much faci­ Veenkolonien may exceed this maximum value. litated by the shape of the pit and a littie wind. Seen from above, the base is mostly round to oval; Where the subsoil is suitable for digging a hearth-pit in the latter case no dominant orientation could be (as with the unpodzolised coversands in the Meso­ found. What form the mouth of the pit possesses lithic) its construction is ajob that takes only a short could not be recorded, as this zone is completely while. bleached by the A2 horizon. The colour of the fill of Furthermore, the experimem suggested that it is the pit may be any shade from grey to black, primarily the process of burning that determines depending on the amount of charcoal left, often how much charcoal is left rather than the extent of with a greenish tinge. The lower part of the pit leaching or clearance of the hearth. The fires that usually contrasts sharply with the reddish-yellow of were extinguished will have more charcoal re­ the surrounding B3 horizon, although infiltration maining than those which were left to burn out. of humic matter is often lacking in the immediate b) The fi ll of the pit. The majority of the Meso­ vicinity of the pit. The concentration of charcoal is lithic hearth-pits have a very dark-coloured base, as densest at the base, gradually fading out upwards. generally there is a lot of charcoal left. The charcoal In the charcoal no horizons were observed; the fill at may be encountered in fragments that retain their the base of the hearth-pits offers a very homo­ botanical identity, such as bits of branches, or geneous picture. consists of strongly fragmented material and even To collect information about the effort required charcoat dust the botanical determination of which to dig a Mesolithic hearth-pit and about the effect of is not possible. However, not all the hearth-pits are burning a fire in it, experimental hearth-pits were rich in charcoal, some showing a less contrasting, constructed in unpodzolized coversand (Groenen­ grey-cotoured fill with only very fine particles of dijk & Smit, in prep.). Some relevant results are charcoal left. In none of the hearth-pits was any mentioned here. Scooping out an arm-deep shaft la�'er found in the fill, that could be indicative for not only gives a depth of c. half a metre, but multiple use or incomplete c1earance. The objection automatically produces a bowl-shaped base. . Pro­ coulct be raised that one is una ble to recognize vided there is some draught, burning a fire of Pinus, cleared pits in field situations. This may be true 'as af ter a slow start, combines a high temperature with long as the charcoal was completely removed, not low fuel consumption due to the construction of the only from the pit but also from around the opening shaft. Widening the shaft makes the fuel run out of the pit, and subsequently dumped elsewhere.14 faster under the same conditions. Af ter the phase of No clear patches of charcoal around hearth-pits 94 H.A. GROENENDIJK

were observed in cases where the old surface around flint was found in very small quantltles (only the hearth-pits was virtuaIly intact. In fact the fissured and red-coloured flint was registered as opposite was recorded: mere smudges of charcoal 'burnt'; pieces without these characteristics may with vague contours, often observed at the top of however have been in contact with fire too). On the the A2 horizon, hardly ever turned out to be hearth­ whole, the category of 'unburnt flint' sinks into pits when pursued at a deeper level. But even if insignificance in comparison with the quantity of clearance did take place, an impressive number of clearly burnt flint artefacts found in hearth-pits. abandoned hearth-pits full of charcoal remain to be This was to be expected, for only the charcoal bases explained. of the hearth-pits were analysed. The observation Even though clearance is unlikely, multiple use that burnt flint artefacts dominate the lithic mate­ cannot be ruled out, although both phenomena can rial found in hearth-pits (only in one hearth-pit be part of the same activity. Multiple use (i.e. minor fragments of sandstone were found).is a first interrupted use) is much more difficult to demon­ step in the interpretation of the role of these hearth­ strate; the absence of stratification in the charcoal pits within a Mesolithic site; this point will be does not necessarily prove single use. Single use on worked out in the folIowing. the other hand may have extended to a length of Identifying and counting burnt flints is com­ days. The large quantity of charcoal left as a plicated and precarious because of fragmentation homogeneous layer in many of the hearth-pits, is an through contact with fire. Nevertheless the majority argument for the extended (single) use hypothesis. of the burnt flints could be identified; flakes and The hearths will then have been continually refuel­ blades dominate, but cores are also found. This led. underlines the waste character of the flint found in We do not know if the process of eventually the hearth-pits. Sporadically retouched tools occur: refilling the pits with coversand was a natural one or two points, two scrapers, one backed blade and one was done deliberately. Mostly the sand in the upper retouched blade were recorded altogether.15 Pro b­ part of the shaft appeared sterile, but it might also ably these retouched artefacts must be considered as contai n fine particles of charcoal as well as unburnt waste too, although the damage caused by burning flint artefacts. In the case of unburnt flint occurring does not allow a detail ed examination. The small­ in the fill above the charcoal base one is inclined to sized waste material must have got into the fire think of a gradual refilling of the pit af ter the fire accidentally. VirtuaIly all flint found in the hearth­ burnt out. pits was heavily burnt; both a high temperature and prolonged exposure to the fire may be responsibie 2.2.5. Contents for this. Obviously flint working was done near :>0lJle of the hearth-pits while they were in use. For the analysis of the contents the excavated bases Theoretically, the presence of flint in a hearth-pit of the hearth-pits were wet-sieved; a sieve of 2 mm cquld for a part be explained as flints, falling into mesh was used. Figure 6a shows what was the pit from the dug-out soil aiready containing encountered in the 111 hearth-pits from the Veen­ them. However, such a situation would only ac­ kolonien (the few analysed hearths from beyond the count' for very low numbers of burnt flint present in study area displayed no obvious differences). Apart hearth-pits. from charcoal, burnt flint is the main category With figures 6b-d we aim to demonstrate that not observed; occasionally what we think to be unburnt just the total ofhearth-pits containing burnt flint in

82 17

l C]

2 f72Z1

3 �

Q. all heorlh-pils IN,ttl) b. SSIN,IS) C. NP 3 IN,43) d. S 51 I Nd5)

Fig. 6a-d. Diagrams of the contents of Mesolithic hearth-pits from the Veenkolonien. Legend: l. charcoal only; 2. containing charcoal and bumt flint; 3. containing charcoal, bumt flint and unbumt flint.The numbers represent the totals of all hearth-pits concemed. 6a: all III analysed hearth-pits (from 6 sites altogether); 6b, 6c and 6d: the contents of lhe sites S6, NP3 and S5 1; only siles with a sufficient number of hearth-pits have been selected for this purpose. Mesolithic hearth-pits in the Veenkolonien 95 addition to charcoal may vary from site to site, but working area next to these high-scoring hearth-pits; also that the number of bumt flints per hearth-pit implicitly this presupposes contemporaneity. This may vary with the proportion of hearth-pits con­ could still not be substantiated however: an attempt taining bumt flint from any single site: in the 2 to refit the flint material did not yield positive hearth-pits containing bumt flint among the 16 on results and none of the hearths in question has been site S6, only 3 bumt pieces were counted altogether. radiocarbon dated so far. But another observation Site S51 represents the other extreme; 17 heatth-pits may be of importance. out of 35 contained bumt flint, with an average of Among the bumt flint artefacts from the three 12 pieces per hearth. This site, with the highest rate hearth-pits, the varying extent of exposure to fire is of hearth-pits containing bumt flint, allows a more seen to produce a bipartition: both white-bumt, detailed examination. The meaning of the cluster of fissured pieces and fissured pieces that have more or hearth-pits on this site, presumed to be contempo­ less kept their original transparency occur. With raneous, has aiready been discussed (section 2:2.3). figure 7 it ean be demonstrated that: The bumt flint found in them, occutring In low a) Whitened, fissured pieces predominate in quantities, is distributed over the hearth-pits at the number over the 'fissured-only' pieees, in spite of westem edge of the cluster. If, as suggested, the the fact that whitened, fissured pieces are the most hearth-pits really functioned with only a short time heavily burnt and thus would have had less chance interval (or even simultaneously), flintworking may of surviving as complete artefacts. Moreover flints have taken place to the west of this cluster. The of this category surpass the 'fissured-only' pieces in waste material then just fell into the hea.rthsthat lay size. This provides an argument for the idea that nearby. more than one flint-working session is reflected in The same applies to an area in the W part of the the contents of the three hearth-pits. site. There, three hearth-pits lying no more than six b) Dimensions of under 10 mm are the most metres apart (though qivided by an unexcavated numerous with both categories, emphasizing the strip), contained 116, 20 and 15 pieces of bumt flint waste character of the flint in both cases (in faet only artefacts against 5, 5 and 6 pieces for adjacent the scraper is questionable as such). hearth-pits. Again, it is tempting to locate a flint

JO

I C s

25

20

10 Fig. 7. The burnt fl int from the three high-scoring hearth-pits on site S51 taken together, showing the differ­ ence in length and width as well as in number of whitened, fissured flints and 'fissured-only' flints (the diffe r­ lI1l ence is emphasized by contouring the � values ofboth categories, s = scraper, c = core). Measured are the complete artefacts; scale is in mm. Legend: l. width whitened, fissured flint; 2. 'fissured- only' flint. 10 15 20 25 JO 96 H.A. GROENENDIJK

With a bipartition in the extent of buming as well as A material category from Mesolithic hearth-pits in the size and number of the pieces from both elsewhere, not encountered in the Veenkolonien, is groups, we wish to argue that on two separate bone. The presence of human calcined bone in occasions flint waste got into the hearth-pits while Dalfsen, province of Overijssel and Oirschot, pro­ they were in use: one batch of flint was heavily vince of Noord-Brabant (Verlinde, 1974; NewelI, bumt, the other was bumt to a lesser extent. The 1975; NewelI et al., 1979; Arts & Hoogland, 1987) is effect of the position of the flint within the fire with clearly connected with Mesolithic hearth-pits, How­ respect to temperature (referring to hotter or'cooler ever, for the interpretation af intentional human areas in the fire) can be negleeted, as the type of cremation taking place in such hearth-pits, the hearth - given a very concentrated and well­ evidence in both cases is rather thin. insulated fire that moreover may bum for a long Stones (as part of a stone structure within the time (see below) - suggests a uniform process of hearth ar as cooking-stones) sporadically occur in fissuring and discolouring of the flint; the only MesolithlG hearth-pits, although cooking-stones are difference would be at what phase in the use of the known in Mesolithic contexts (see e.g. Arts & hearth-pit (right at the start or at the end) the flint Hoogland, 1987 for Oirschot V; and Groenendijk, came into contact with the fire. If this hypothesis is 1986 for NP3). Same af the flint 'cooking-stones' right, multiple use, or perhaps rather an extended from NP3 (location: see fig. 2) showed clear traces single use of these hearth-pits, long enough to cover of contact with fire, others looked unaltered, but two sessions of flint working, is indicated. The none of them were as heavily bumt as the majority initial point of discussion, the possibie contempo­ of flints found in the hearth-pits. raneity, may even apply to this group of three A distinction therefore should probably be made hearth-pits. The levelling af the Mesolithic surface in the application af fire with respect to cooking­ at this spot however makes a comparison with the stones against the flint artefacts that were casually actual flint assemblage at the surface impossible, thrown/accidentally fell into the fire. Future radiocarbon dates may support or refute The most important conclusions from the sec­ t�is view. tions 2.2.4 and 2.2.5 are: In the charcoal, Pinus is dominant in the Pre­ a) The hearth-pits did not include structural boreal and Boreal hearth-pits from the Veenkolo­ elements such as stone·constructions. nien. Quercus emerges in small quantities in the b) In most af them only charcoal was observed, Boreal and increases only at the end of the Boreal carbonized remains of vegetable foodstuffs or cal­ and the beginning of the Atlantic. In the charcoal cined bone seldom occur. other wood species such as Ulmus and Populus c) Some hearth-pits lay next to flint-working appear towards the Atlantic.16 This may reflect the areas, as the flint material found in the hearth-pits is changing supply of firewood, although Pinus pre­ waste. dominates by far throughout the Mesolithic oc­ d) This flint working apparently was done while cupation of the Veen kolonien. Of this species often the hearths were in use, indicating extended use of large chunks af charcoal (of branches with a the fire; a functional relationship between the use of diameter of up to 5 cm) are encountered. We suspect hearth-pits and flint working may exist, but cannot a relationship between the preservation of these be demonstrated. large lumps and the abundance of Pinus charcoal in the hearth-pits, Pinus must have been a highly 2.2.6. Restrietion in time appreciated fuel for many centuries because of its charcoal that keeps glowing for a long time; this The earliest Dutch Mesolithic radiocarbon samples indicates a preference for a slow, smouldering fire. date from af ter roughly 9500 BP (Lanting & Mook, Especially in the Boreal Pinus was abundant. It is a 1977; in prep.). This is in agreement with the first tree under which broken branches can be easily appearance of Mesolithic occupation in adjacent collected. areas, such as Duvensee, Germany, which has A carbonized shell of hazelnut was found in one ample radiocarbon dates (Bokelmann, 1971). These hearth-pit only, although a more frequent occur­ dates, slightly earlier than the Dutch ones, do not rence could be expected because af the presence of concem hearths, but culture layers and peat sedi­ scattered carbonized nutshells on several of the ments. The same goes for Denmark, where dated excavated sites in the Veenkolonien. Storage of hearths even are relatively young (Tauber, 1973). hazelnuts may have been the case with a pit in the For the beginning of the Mesolithioih the Nether­ shape of the hearth-pits under discussion, as it Iands, Ganting and Mook follow Newell's chrono­ seerned full of complete, carbonized hazelnuts near logy (NewelI, 1973; 1975a). It is striking to see that Stadskanaal (an observation made in 1952 by J.H, these earliest Dutch Mesolithic samples are taken Keizer, Groningen). The Mesolithic, human origin from hearths, morphologically belonging to the , of this pit seems certain. type of 'Mesolithic hearth-pits .17 Mesolithic hearth-piJs in the Veenkolonien 97

Waterbolk suggests that the absence of radio­ possibly used in the same tradition) was found here carbon dates before c. 9400 BP in the Netherlands by chance. may partially be explained by the combustibles The same problem may be involved at the De available and their archaeological properties (Wa­ Leien-Wartena settlement Bergumermeer S64, pro­ terbolk, 1985). According to this argument, fire­ vince of Friesland (Aarts, in prep.; NewelI, 1980; wood of Betula- a predominant tree throughout the Lanting & Mook, in prep.; and RR. NewelI, pers. first half of the Preboreal - would leave little or no comm.). A diversity in hearth-pits was noticed both charcoal, contrary to Pinus. Thus we would not be on the site and after laboratory analysis; yet a group able to recognize the earliest Mesolithic 'Betula' of 'Mesolithic hearth-pits' could be isolated. A part hearths. Although the preservation of charcoal will of the radiocarbon dates from this group proved to primari ly depend on the burning process, as dis­ be younger, namely from Early/Middle Neolithic to cussed in section 2.2.4, we think there is an argu­ lron Age. ment against Waterbolk's considerations. Even if These examples may show that setting cultural we accept that Pinus was absent before c. 9400 BP, boundaries for the use of 'Mesolithic hearth-pits' Betula is suitable for kindling an op en fire only, but does not make sense, unless we kno w more about is no good for keeping a slow fire going as is Pinus their use and function within a site. Partly due to the (or rather, Pinus charcoal). This makes us turn to fact that most Neolithic excavations focus on another of Waterbolk's suggested explanations for house-plans and funerary rites, the type of the the lack of dates earlier than c. 9400 BP: the practice 'Mesolithic hearth-pit' gradually slips out of view. of digging pits to construct hearths did not develop The tradition of using this apparently successful until that time. It was with the first appearance of type of hearth may however have survived long Mesolithic man in these regions, that 'Mesolithic af ter its 'Mesolithic' application. hearth-pits' were introduced. From the samples of hearth-pits delivered at the Groningen Centre for Isotope Research, an increase 3. TOWARDS A FUNCTIONAL INTERPRET A­ of sites in the Netherlands from the Early Mesolithic TION onwards (c. 8700-8200 BP) can be noticed (Lanting & Mook, 1977). This tendency will main ly reflect 3. 1. Outside the domestic area the traditional strategy of investigation, which was It has been argued that the Mesolithic hearth-pits hardly directed at exploring the earliest Mesolithic discussed here probably reflect only one type of sites (R.R. NewelI, pers. comm.). The radiocarbon­ hearth present on Mesolithic sites in the Pleistocene dated hearth-pits from the Veenkolonien suggest an coversands of the Northwest-European Plain, in increase in the number of sites, which started spite of tl1eir numerosity and the. great number of around 9000 BP (= Basal Mesolithic); this is slightly sites that include these hearth-pits. Surface hearths, earlier than in the Netherlands as a whole. How­ especiaIly those lacking resistant structures like ever, only twelve dates are available at present, a stones, -exposed to the weather, may have been number still toa small to draw conclusions. easily eroded by the wind and become 'invisible' in From c. 6300/6200 BP onwards Mesolithic oc­ the archaeological record. Of their actual presence cupation in the Netherlands can no longer be traced in the study area hardly anything is known, nor (that is, younger samples have not been delivered at have they been described from other Mesolithic the Centre for Isotope Research). The reason be­ sites in the Netherlands. hind this is not yet clear; it does not seem reasonable The nearest examples of Mesolithic hearths of a to assume that the Mesolithic just disappeared from different type have COlne to light at Duvensee, the radiocarbon screen and proceeded without northern Germany (Bokelmann, 197 1; Bokelmann leaving any traces (J.N. Lanting, pers. comm.). et al., 1985; Bokelmann, 1986). Their central posi­ Be that as it may, it is definitely wrong to connect tion in 'domestic' contexts (viz. associated with 'Mesolithic hearth-pits' with the Mesolithic only. bark floors and flint artefacts concentrated around Near the study area, in the gemeente Stadskanaal, a or near the hearths), as well as their construction as charcoal-stained pit, which morphologically can be surface hearths and their multiple use (as indicated ranked among the 'Mesolithic hearth-pits' was by the clearing zone around the 'big' hearth from found underneath a Late Neolithic/Middle Bronze Wohnplatz 13), rank this hearth-type among what Age barrow; any association with the barrow can be might be called 'domestic hearths', in contrast with excluded. Determination of the charcoal by W.A. the hearth-pits under discussion and in spite of the Casparie, B.A.I. (fast-grown Quercus) aiready rai­ fact that the excavator explicitly suggests an open­ sed doubt as to its Mesolithic origin; besides the air location for the 'big' hearth of Wohnpfatz 13 radiocarbon age turned out to be Middle Neolithic (Bokelmann, 1985). (4750±40 BP, GrN- 15315). Thus a Neolithic With the Early Boreal Duvensee hearths in mind, hearth-pit, similar to the Mesolithic ones (and we consider the lack of 'domestic' hearths in the 98 H.A. GROENENDIJK

Dutch Me!iolithicso far to be due to an observation forrned during several visits to the camp site this hiatus in connection with the very different condi­ argument is valid, for it is reasonable to expect that tions of preservation. Nevertheless the fact that the in the limited space of a domestic area a hearth Duvensee hearths were built on top of wet organic would be used again and again instead of new deposits should not be neglected. It is an open adjacent ones being constructed. question to what ex ten t the subsoil determined the Re-occupation of a camp site apparently did not type of hearth used. entail the clearance of hearth-pits. It may have been We wish to put forward here that Mesolithic the ease of digging a new pit in the loose coversand hearth-pits, other than 'domestic' hearths, lay out­ compared to clearing an old pit why the latter was side, or at least not at the centre of, a domestic area avoided. Instead, old pits may have been filled up (hut, tent or living floor). deliberately for other, e.g. hygienic reasons. This Several arguments support this hypothesis. In the raises the question of whether in the case of a first place, with the Mesolithic hearth-pits we miss seasonal re-occupation of the camp sites in the elements that characterize hearths which are part or Veen kolonien, the remains of abandoned domestic the centre of a domestic unit, such as a peripheral hearths within domestic units were incorporated zone with charcoal (clearing zone) and worked flint, in to the new camp site, including clearance of old indicating activities around the fire that include flint hearths. This question cannot be answered now working (in the acid coversands other materials through lack of appropriate analyses of Mesolithic such as bone have only sporadically been pre­ sites, but would be of great importance to testing the served). Besides, 'domestic' hearth-pits as a rule suggested distinction of 'domestic' and 'non-do­ would have contained flint waste, and not inci­ mestic' hearths in the Mesolithic. dentally as they do now. Nor were any structural elements observed, as in the case of Late Palaeo­ 3.2. Functional aspects lithic, presumably domestic hearths, such as a stone lining or stones in the hearth-pit itself. The shape of The specific construction and the prevalent use of the hearth-pit will not have necessitated any ad­ Pinus firewood over thousands of years, supposes a ditional stone structure. firm tradition in the use of hearth-pits that cor­ Moreover, Mesolithic hearth-pits are not mor­ respond with a slow-burning fire. phologically suited to use in a domestic area, where Our hearth-pit experiment showed that a short one would expect a fire burning at or near the phase of open, fierce fire could soon be followed by surface, well accessibIe and able to radiate heat. The a long phase of actual use with live embers, the form of a hearth-pit is fundamentally different and duration of which depended on the strength of the produces an encased fire. The relatively deep shaft wind. The experimental roasting of food (in our does not fit in with the use in a domestic unit case: hazelnuts) needed a smouldering fire and in (presumably protected against the winds somehow, fact only the surrounding sand thus heated was while hearth-pits need at least a littie wind). It found useful for roasting. Bokelmann et al. (1985) hardly radiates any heat. One might perhaps put it came to the same conclusion with respect to roast­ the other way around: because of their location in ing technique with the 'big' hearth of Wohnplatz 13. the open air the fires were protected against toa With these remarks on experimental roasting, based much wind by burning them in deep pits. In this on the fact that carbonized hazelnut shelIs are light, the separate location is likely to have been tied sometimes encountered in Mesolithic hearths, a to peripheral activities. first interpretation in terms of use is aiready given. For Mesolithic hearths that include (some of)the Not only does the taste of hazelnuts improve by 'domestic' elements we thus have to turn to Duven­ roasting, it makes conservation possibie till the next see. The Late Palaeolithic hearths listed in section harvest. The hazel does not emerge before the end of 2.1 may in the same sense be considered as 'do­ the Preboreal, so that the idea of storing hazelnuts mestic' , i.e. associated with a hut or tent structure. could only have come up since that time. Never­ They are surface hearths, include flint working theless the roasting of hazelnuts is not associated around the hearth, stone structures and clearing with hearth-pits only, as is clearly shown by the zones with charcoal. For a 'domestic'/'non-do­ Duvensee surface hearths. Here toa it could be mestic' comparison to test this hypothesis, Late demonstrated that consumption of unroasted hazel­ Palaeolithic data unfortunately are lacking for the nuts must have been the practice as well (see the Northwest-European Plain. 'small' hearth, Wo hnplatz 13). Secondly, the observed contemporaneous cluste­ Mesblithic hearth-pits however do not seem to be ring of hearth-pits suggests a location outside the 'designed' for food preparation exclusively. The domestic area as well. It is the extent of the cluster smoking and drying of non-food objects (e.g. hide areas as well as the scarcity of transections that smoking and the drying of animal and vegetaele point in this direction. Even if the clusters were foodstuffs) can be mentioned as another possibie Mesolithic hearth-pits in the Veenkolonien 99 application. Further, the roasting or meat (and sray unattended for a while, food being processed smoking of meat as a preservative?) must be serious­ above or in it; only reloading with fresh firewood ly taken into consideration. Possibly the size of the entails an interruption as the wood has to pass shaft was varied according to the application. through the stage of forming charcoal again. The Whether the object was placed above the fire, into success of this type of hearth is evident from the the glowing embers or into the underlying heated extremely high numbers encountered in the North­ sand, in all cases .the great' advantage over a surface west-European Plain and the apparently long life of hearth is a highly controllable fire, giving a mode­ this hearth-pit tradition. rate, even heat for an extended period. As a further Does the Mesolithic hearth-pit constitute an technical advantage procedures can be facilitated innovation, emerging in the second half of the by placing stakes over the fire into the wall of the Preboreal? With a glance at the excavated Late shaft at varying distances from the fire. The techno­ Palaeolithic hearths from the Northwest-European logy of storing seasonal resources may have de­ Plain one is inclined to think so. Our tentative veloped parallel to roasting for immediate use. comparison with Late Palaeolithic surface hearths The archaeological evidence in the Netherlands is based on incommensurable data however. Yet offers little perspective on other explanatory ap­ some indirect arguments in favour of the emergence plication than food processing. Nevertheless, a of a new type can be mentioned. Let us therefore put continuous fire at a camp site actually may have the question the other way around: were environ­ stimulated multiple applications. Indeed it may mental conditions in the Late Palaeolithic favour­ have been necessary to have more than one hearth able for using hearth-pits? Climatic change for one in use at the same time, which could explain the thing is unlikely to have altered the coversand presence of clusters. This should be a point of subsoil in a way to affect the digging of pits for fires. further investigation; detailed study of the contents The composition of the topsoil and the vegetation of Mesolithic hearth-pits in other areas, and in thc on the camp location are liable to have played a Veenkolonien a further examination of the burnt more. decisive role.ls These circumstances were not flint artefacts from hearth-pits and the question of unfavourable for the use of hearth-pits in the Late whether flint working can be associated with them, Palaeolithic. The absence however of plant food may offer clues to this suggested multiple function. resources that would have stimulated food pro­ Activities connected with the use of fire, which cessing for storage may serve as an argument to perhaps needed more space than was available suggest a Mesolithic origin. With food processing within the domestic unit, or were thought toa messy still regarded as the main objective of the use of Ol' genera ted toa much smoke to be perforrned hearth-pits, the Mesolithic environment.offered an within, may have played their role in demarcating extending range of small-sized edible items, both of an area of specialization involving the use of fire floral and faunal origin, perhaps consistent with outside the domestic area. We can not corroborate fires in narrow pits. This implies that the size (or this supposition with archaeological evidence. The more specifically the width of the shaft) of the possibly relevant flint måterial has not been spatia 1- hearth-pit may depend on the type and size (of the ly analysed until now, and the study of micro-wear pieces) of the food prepared. Hearth-pits may thus traces on flint artefacts from the Veenkolonien has reflect changing environmental conditions, in that proved unsuccessful, considerable mechanical wear their form may follow the function of the hearth. af ter site abandonment being the reason (informa­ A continuation of Late Palaeolithic traditions in tion provided by A.L. van Gijn, Instituut voor the use of fire (i.e. 'domestic' hearths) might be Prehistorie, Leiden). present in the Mesolithic Duvensee hearths. For the investigated area we lack such information. The absence of hearth-pits in a Late Palaeolithic context 3.3. Conclusion: a new type of hearth? is suggested by the archaeological record and is EspeciaIly for food processing a deep, shaft-like pit supported by indirect arguments of an environ­ with a smouldering fire at its base constitutes an mental nature. Seen from this point of view the enormous improvement over surface hearths. A quest for the oldest 'Mesolithic hearth-pit' has surface hearth on the whole makes for a rather become relevant. inefficient use of fire, combining high fue! consump­ Concerning the role of hearth-pits within the tion with low return. In surface hearths heat can range of activities on Mesolithic camp sites some only be retained by means of (a structure of) stones tentative ide as were put forward, especiaIly with and always needs protection against weather in­ regard to their location outside the domestic unit(s) fluences. Not so hearth-pits. Their dependence on proper, as well as to their extended burning for atmospherical conditions basically will be confined special purposes. Finding out to what extent these to the starting phase, when a bit of wind is required. reconstructions are valid must be one of the ob­ Arriving at the stage of glowing embers, the fire can jectives of future study. 100 H.A. GROENENDIJK

4. ACKNOWLEDGEMENTS 5. The interpretation of the shape of the hearth is on the account of the present au thor; the original publication merely contains a section drawing of il. The discussion about Mesolithic hearth-pits actual­ 6. This dating corresponds with that ofthe Usselo layer in the ly started when in 1984 excavations at several Netherlands, although for an Ahrensburgian hearth a Mesolithic sites yielded an impressive number of younger date would be expected. hearth-pits. When in the text the 'we' form is used, 7. This interpretation ofthe zones is the present author's view. this is done in recognition of the many discussions I 8. For instance, at the Magdalenian site of Pincevent ih the Paris Basin 'satellite hearths' were encountered at some had with my fieldwork comrade John Smit (Wil­ distance from the 'domestic hearths' which they were dervank), who critica,lly followed the subject cul­ associated with. They differed from the latter in their minating in a hearth-pit experiment (1986); but I morphology and function (Julien, 1984). remain solely responsibie for the content of this 9. On the investigated sites in the Veenkolonien other features are possibly associated with fires. These are: (a) heteroge­ article. During the writing of this paper J.N. neous, shallow 'pits', presumably man-made, with smudges Lanting and R.R. NewelI (B.A.1. and B.A.I./SUNY, of charcoal; (b) high concentrations of burnt flint artefacts Binghamton) kept me from the worst slip-ups. I without hearth-pits. An example is site NP3 (fig. 2) with wish to thank the Centre for Isotope Research, flint concentrations in the extreme NEand SW, both with a clear centre and associated with charcoal, but in different Groningen University, for rapidly providing me ways: the NE concentration with a 1.5 m wide depression with some radiocarbon dates, and W.A. Casparie including a central pit specked with particles of charcoal, (B.A. 1.) for the determination of the charcoal. The the SW concentration with particles ofcharcoal around the English tex t was corrected and in parts translated by level of the old surface in a density that exceeds the general Xandra Bardet (Groningen). degree of charcoal pollution at the site. IO. The homogeneous nature of Mesolithic hearth-pits is contested by R.R. NewelI on the grounds of his analyticai data concerning the total amount of ground features that 5. NOTES may occur in a single site: by means of this ground feature analysis, including all hearth-pits, he recognizes a dif­ ferential use of fire within settlements (R.R. NeweII, pers. in the l. In his survey of Mesolithic and Neolithic settlements comm.; Aarts, in prep.; see also NeweII, 1975; 1980). In view draws attentIOn to northernNetherlands, H.T. Waterbolk ofthe fact that so fa r no complete Mesolithic sites have been 1985). the occurrence of Mesolithic hearth-pits (Waterbolk, excavated in the Veenkolonien, such an analysis cannot be prompted me to Some of his remarks on the subject have perforrned here. However, there has been extensive co­ in greater detail, and to take the investigate this subject . verage in the form offield observations and the number of . as the many hearth-pits uncovered in the VeenkoloOlen hearths tllUS found is considerable. Therefore a different starting-point for this contribution. . approach has been opted for in this case. From among the mto the Meso­ 2. The geomorphological aspects of research ground fe atures within the investigated area only the b the author, lithic occupation are part of a study prepared 7 Mesolithic hearth-pits immediately recognizable in the field of the dealing with the landscape and settlement hlstory as 'classical' specimens (see section 2.2.4) were selected, and Oost-Gronm­ redevelopment area 'Herinrichtingsgebied these have been analysed on a number of relevant points The Veen­ gen en de Gronings-Drentse Veenkolonien'. (see section 2.2.5), ItS geomorpho- kolonien are part of this area (for some of II. An example of the latter is Havelte-De Doeze, province of logical aspects see Groenendijk, 1988). . . Drenthe (Price, 1975). There the few radiocarbon samples mformatlon on the 3. On the site NP3 we gathered detailed dated so far from six adjacent, but separate Mesolithic sites flint concentration vertical distribution of flint artefacts. A were not in accordance with the typological assignments of of raised bog. lay partly undisturbed under a residual layer the flint tools fo und in them. to the B2 honzon From the top ofthecoversand dune down 12. Once contemporaneous or even simultaneous use of Meso­ of artefacts was of the podzol profile the distribution lithic hearth-pits on one site can be deduced from the and Size. The recorded with reference to their frequency archaeological c:vidence, these clusters could act as a cross­ 0-5 cm (A2 with fo lIowing depths are mean values: from check for the application of the calibration curve for the in the upper 2 cm) secondary infiltration of humic matter period in question. In that case all hearth-pits of a cJuster from 5-7 cm only some splinters of flint wereencountered; should be radiocarbon dated. For the clusters under they remamed (A2) the number of artefacts increased but discussion this has not yet been done. transItIOn B2) a under I cm in size; from 7-9 cm (A2 down to 13. T.D. Price has suggested that infra-Ted photography could and from sudden increase in number and size was observed be a help in distinguishing othenvise invisible mi nor and lower) only the top of the B2 downwards (9 cm organic particles in the dark brown B2 horizon or the both larger sporadically artefacts were found, though bleached A2 horizon of fe atures in podzolized coversands. Thls distributIOn pieces and splinters occurred at this level. Thus the shape of a feature may be reconstructed up to the after in our opinion indicates a long period ofbloturbatIOn original surface level (Price, 1975). We did not apply this apparently the camp site was abandoned; sedimentation method in the Veenkolonien. did not take place. 14. For a charcoal dump we possess an indication at site WVIO, 4. An exception must be made for the Mesolithic site S6, gemeellIe Veendam, where carbonized branches of Pin I/S situated near a stream valley (the Pagediep), part of a were fo und without evidence of a hearth-pit (Groenendijk drainage system that is likely to have become frustrated not & Smit, 1984/85). until the second millennium BC (unpublished research). In 15. Radiocarbon dating of hearth-pits in which tools are found this valley peat growth started much later than elsewhere in could give a terminus Gnle quem for the artefact type. This I the Veenkolonien because of a better natura dramage. has not yet been systematically studied; the tools found so Besides Mesolithic flint material and Mesolithic hearth­ far do not allow a narrow chronological attribution. pits, further investigation also brought to light Neolithic 16. Determinations and provisional conclusion by W.A. Cas­ flintartefacts. parie (B.A.!.). Mesolithic hearth-pits in the Veenkolonien 101

17. We ean add an early radiocarbon date from sue h a hearth­ KROMER, B., M. RHEIN, M. BRUNS, H. SCHOCH-FI­ pit in the study area (gemeelIte Hoogezand-Sappemeer) of SCHER & K.O. MUNNICH, 1986. Radiocarbon calibration 9470±70 BP (GrN-11996). data for the 6th to the 8th millennia BC Radiocarbon 28, pp. 18. The same may apply to wet organic sediments as . at 954-960. Duvensee. Using hearth-pits iil the way as was current in LANTING, J.N. & W.G. MOOK, 1977. Th e pre- and proto­ coversands in a wet subsoil is likely to fa il. In this context, history of the Netherlands in temlS of radiocarbon da tes. further stucty of Mesolithic hearth-pits dug into different Groningen. . soil types would be ve ry" useful. LANTING, J.N. & W.G. MOOK, in prep. Th e pre- and protoilistory of the Netherlands in terllls of radiocarbon da tes. 2nd updated ed. Groningen. MOLEMA, J., 1988. De vuursteenassemblages van jong-paleoli­ 6. REFERENCES tische en mesolithische kampplaatsen in het Groninger veen koloniale gebied. Groningen (unpublished typescript). NEWELL, R.R., 1973. The Post-Glacial adaptations of the AARTS, M.E.N., in prep. A statisticai analysis of the Ber­ indigenous population of the Northwest European Plain. In: gumermeer S-64 ground . features: defining natural pro­ S.K. Kozlowski (ed.), Th e Mesolithic inEuro pe. Warsaw, pp. venience units in a Mesolithic settlement. 399-440. ARTS, N., 1984. Waubach. A Late Upper Palaeolithic/Meso­ NEWELL, R.R., 1975a. Mesolithicum. In: G.J. Venvers (ed.), lithic lithie raw material procurement site in Limburg, the Noord-Brabant inpre- enprotohistorie. Oosterhout, pp. 39-54. Netherlands. He/illium 24, pp. 209-220. NEWELL, R,R., 1975b. A proposed attribute analysis of ARTS, N., 1987. Vroegmesolithische nederzettingssporen en archaeological ground fe atures: an expansion of the auto­ twee versierde hertshoornen artefakten uit het Maaskant­ matie artifact registration system. In: S. Laflin (ed.), Compu­ gebied bij 's-Hertogenbosch. Braballts Heem I, pp. 2-22. ter app/ications in archaeology. Proceedings of the anlllwl ARTS, N. & M. HOOGLAND, 1987. A Mesolithic settlement conference organized by the Computer Centre University of area with a human cremation grave at Oirschot V, municipa­ Birmingham. Birmingham, pp. 69-80. lity of Best, The Netherlands. Helillium 27, pp. 172- 189. NEWELL, R.R., 1980. Mesolithic dwelling structures: faet and ASSENDORP, J.J., 1985. Ein Fenster in die Vergangenheit. In: fa ntasy. Veroffent/ichungen des Museums flir Ur- lind Fnih­ AlIsgrabulIgell 1979-1984 = Berichte zur Denkmalpflege in ( geschichte Potsdam 14-15, pp. 235-284. Niedersachsen, Beiheft I). Stuttgart, pp. 78-80. NEWELL, R.R., T.S. CONSTANDSE-WESTERMANN & C BOKELMANN, K., 1971. Duvensee, ein Wohnplatz des Meso­ MEIKLEJOHN, 1979. The skeletal remains of Mesolithic lithikums in Schleswig-Holstein, und die Duvenseegruppe. man in Western Europe: an evaluative catalogue. Journal of Offa 28, pp. 5-26. HlIman Evolution 8, pp. 5-228. BOKELMANN, K., 1986. Rast unter Biiumen. Ein ephemerer PADDA YY A, K:, 1971. The Late Palaeolithic of the Nether­ mesolithischer Lagerplatz aus dem Duvenseer Moor. Offa Iands - a review. He/iniulll II, pp. 257-270. 43, pp. 149-163. PERLES, C, 1977. Pn?histoire dufeu. Paris. BOKELMANN, K., F.R. AVERDIECK & H. WILLKOMM, PRICE II, T.D., 1975. Meso/ithic selllement systems in the 1985. Duvensee, Wohnplatz 13. Ojfa 42, pp. 13-33. Netherlands. Diss. Ann Arbor. BOSINSKI, G., 1979. Die Ausgrabungen ill GOllnersdOl! 1968- SCHWABEDISSEN, H., 1954. Die Federmesser- Gruppen des 1976 ulld die Siedlungsbefl lnde der Grabung 1968 (= Der nordlVest-europtiischenFla ch/andes. Zur Ausbreitllng des Sptit­ Magdalenien-Fundplatz Gbnnersdorf 3). Wiesbaden. Magdalenien (= Vor- und friihgeschichtliche Untersuchun­ BRINCH PETERSEN, E., 1973. A survey of the Late Palaeo­ gen aus dem Schleswig-Holsteinischen Landesmuseum f�r lithic and Mesolithic of Denmark. In: S.K. Kozlowski (ed.), Vor- und Friihgeschichte in Schleswig und dem Institut fur Th e Meso/ithic ill ElIrope. Warsaw, pp. 77- 127. Ur-und Friihgeschichte der Universitiit Kiel, Neue Folge 9). BURDUKIEWICZ, J.M., 1986. Th e Late Pleistocelle sIwuidered Neumiinster. poillt assemb/ages ill Western Europe. Leiden. STAPERT, 0., 1970. Zwei Fundplatze vom Ubergang zwischen CAMPBELL, J.B., 1977. Th e Upp er Palaeo/ithic of BritaiII. A PaHiolithikum und Mesolithikum in Holland. Archtiologi­ study of mall and lIature in the Late lce Age. Oxford. sches Korrespondenzblall 9, pp. 159- 166. ES, W.A. VAN, H. SARFATIJ & P.J. WOLTERING, 1988. STAPERT, D., 1982. A site of the Hamburg tradition with a Archeologie ill Nederland. De rijkdom vall het bodemarcilief constructed hearth near Oldeholtwolde (provinee of Fries­ Amersfoort/ Amsterdam. land, the Netherlands); first report. Palaeohistoria 24, pp. GROENENDIJK, H.A., 1986. Een mesolithische pleisterplaats 53-89. bij Nieuwe Pekela. BlIl/etill KOllillk/ijke NederIalIdse Oud­ STAPERT, D., 1986. Two findspots of the Hamburgian tradi­ heidkulldige BOlld 85, pp. 133-135. tion in the Netherlands dating from the Early Dryas Stadial: GROENENDIJK, H.A., 1988. Archeologie en landschap, drie stratigraphy. Medede/illgell Werkgroep Tertiaire KlVartaire voorbeelden uit Oost-Groningen. In: Archeologie ell lalld­ Geologie 23, pp. 21-4 1. scllap. Groningen, pp. 9-30. STAPERT, D., J.S. KRIST & A.L. ZANDBERGEN, 1986. GROENENDIJK, H.A. & J.L. SMIT, 1984/85. Een meso­ Oldeholtwolde, a late Hamburgian site in the Netherlands. lithisehe vindplaats te Wildervank. (jrollillgse Volksalmallak, In: Studies ill the Up per Palaeo/ithic of Britain alld NortInvest pp. 131- 145. Europe (= BAR Intern. Series 296). Oxford, pp. 187-226. GROENENDIJK, H.A. & J.L. SMIT, in prep. Experimenteren STAPERT, D: & F. DE VRIES, in prep. The Hamburglan slte of met 'mesolithische haardjes' (working title). Luttenberg, prov: Overijssel, the Netherlands. Palaeohisto­ GRØN, O., 1983. Social behaviour and settlement structure. ria. Prelimlnary results of a distribution analysis on sites of the TAUBER, H., 1973. Copenhagen radiocarbon dates X. Radio­ Maglemose Culture. Journal of Danis!, Archaeology 2, pp. carbon 15, pp. 86-112. 32-42. TESTART, A., 1982. The significance of food storage among HEESTERS, W. & A.M. WOUTERS, 1970. De Tj ongerkultuur hunter-gatherers: residence patterns, population densities, in De Rips (gem. Bakel). Braballts Heem 22, pp. 2-20. and social inequalities. Currellt Allthropology 23, pp. 523-537. JULlEN, M., 1984. l'Usage du feu il Pincevent (Seine-et-Marne, VERLlNDE, A.D., 1974. A Mesolithic settlement with crema­ France). In: H. Berke, J. Hahn & C-J. Kind (eds.), JlIlIg­ tion at Dalfsen. Bericlltell vall ·de Rijksdiellst voor het Oud­ paltiolithische Siedlungsstrukturell ill Europa. Tiibingen, pp. heidkundig Bodemollderzoek 24, pp. 113-1 17. 161-168. 102 H.A. GROENENDIJK

VERLlNOE, A.O., 1982. Mesolithicum. Marienberg, gem. radioearbon evidenee. In: la hrbuch des Bernischell Histo­ Hardenberg. Archeologische Kroniek van Overijssel over rischeIl Museums 63/64, 1983/84 (= Jagen und Sammein, 1980/81 (= Overijsselse Historische Bijdragen 97), pp. 171- Festschrift H.-G. Bandi), pp. 273-28 1. m. WOUTERS, A., 1957. Een nieuwe vindplaats van de Ahrens­ VERMEERSCH, P.M., 1976. De biostratigrafisehe en litho­ burgeultuur onder de gemeente Geldrop. Braballts Heem 9, stratigrafisehe positie van de epipaleolithisehe en meso­ pp. 2- 12. lithisehe industrieen in Laag-Belgie. In: Extrait du Congres WOUTERS, A.M., 1983. De 'Ahrensburg-sites' van Geldrop. PI"I!historique de Frallce, XXe sessioll, Provellce, 1974. Issou­ Archeologische Berichtell 14, pp. 115-127. dun, pp. 66-71. ZOLLER, 0., 1963. Vorlaufiger Berieht Ubereine Rentierstation WATER BOLK, ti.T., 1985. The Mesolithie and Early Neolithie der Hamburger Stufe bei Querenstede, Kreis Ammerland. settlement of the Northern Netherlands in the light of Die KUllde 14, pp. 17-25.