New AMS dates of the Neolithic and Bronze Age ceramics in Estonia: preliminary results and interpretations/ Uued AMS-dateeringud Eesti neoliitilisest ja pronksiaegsest keraamikast. Esialgsed tulemused ja interpreteeringud.

Kriiska, Aivar ; Lavento, Mika ; Peets, Juri 等

Keraamikatupoloogiatel on labi arheoloogiaajaloo olnud oluline osa esiajaloo periodiseeringute ja kronoloogiate koostamisel. Muud muististe dateerimise meetodid ei ole keraamikatupoloogiat tanini asendanud, kuigi viimast on mitrnel pohjusel ka kritiseeritud. Hilisneoliitilised ja noorernad elupaigad on sageli multiperioodsed, kasutatud katkematult voi vaheaegadega mitmel esiajalooperioodil.

Seetottu on neist keraamikale voi ka teistele leidudele kindlat konteksti raske leida. Tapsemaid voi vahemalt tapsustavaid tulemusi annavad tupoloogiad, mis on aga sarnas meetodina ebatapsed, kui tuupe ei voi siduda loodusteaduslike meetoditega saadud dateeringutega.

Tanapaeval on keraamikatupoloogiaid oluliselt korrigeerinud savinoukildude pinnal sailinud vaikestest soestunud orgaanikakogustest (karboniseerunud toidujaanustest) tehtavad AMS- (Accelerator Mass Spectrometry) dateeringud, saadud vanuste kalibreerimine paikeseaastateks. Kui pinnases ei ole parast kultuurkihi ladestumist mingeid erakordseid protsesse toimunud, siis on korbekiht ja savinou uhevanused.

Olles huvitatud voimalusest savinousid vorrelda, tuleb todeda, et keraamika vanusemaarangud on enamasti ikka veel saadud kinnismuististe, savinoukildude leiukonteksti ja vormi- ning ornamendimuutuste pohjal. Probleemiks on samuti see, et mitmed keraamikatupoloogiad on koostatud aastakumneid tagasi. Samas on aga olulisel maaral lisandunud uusi leide. Nii ei ole eri maades kasutatavad tupoloogiad enam uheselt vordluskolblikud. Uhelt poolt on uued leiud ja kontekstid sattunud vastuollu varasemate tupoloogiatega, sundides aga teisalt ka nende aluseid umber vaatama.

Soome, Karjala ja Eesti tekstiilkeraamika uurimisega on selgunud, et C. F. Meinanderi poolt enam kui poole sajandi eest Laanemere idarannikul eristatud ja pronksiajaga dateeritud Sarsa-Tomitsa tuupi keraamika vajab mitmes osas uut maaratlust. Uks selle keskne tunnus--tekstiilivajutus--esineb tegelikult mitmes neoliitilises keraamikaruhmas ja Eesti ning Vene aladel jatkus tekstiilijaljendiga kaetud savinoude kasutamine kuni rauaaja keskpaigani. Kui tekstiilkeraamikaks loetakse kilde tekstiilivajutuse pohjal, katab see mitmed praegu omaette ruhmadena eristatavaid keraamikatuupe (Eestis hiline kammkeraamika, noorkeraamika, varane tekstiilkeraamika ja tekstiilkeraamika).

Samavord huvitav on ka tekstiilivajutistega kaetud keraamika "algupara" kusimus: kas see lahtub traditsioonist, mille alguse voime dateerida ja lokaliseerida uhte piirkonda, voi on pigem tegu ilminguga, mis on "leiutatud" solturnatult mitmes Euroopa piirkonnas?

AMS-dateeringud anavad aga lisateavet ka nende muististe kohta, kust konkreetsed savinoukillud parinevad, ning loomulikult on uued vanusemaarangud olulised ka tekstiiliajaloo seisukohalt.

Eesti tekstiilijaljenditega keraamika dateerimisprojekti, mille esialgsed tulemused on kaesolevas artiklis avaldatud, peamiseks eesmargiks on luua alus Eestist leitud tekstiilivajutistega keraamika AMS-dateeringutese pohinevale kronoloogiale. Lahtekohaks oli algselt 12 proovist koosnev valim vanimatest ja probleemsetest Eesti asulakohtadest leitud savinoukildudest. Koigilt keraamikakildudelt voetud korbekihtide proovides ei olnud aga dateeringute tegemiseks piisavalt susinikku, mistottu taiendati valimit hiljem uute proovidega. Kui korbekihti ei olnud tekstiilivajutistega savinoukildudel sailinud, siis voeti erandina proov samast asulakohast leitud teist tuupi keraamikalt. Kaesoleva artikli kirjutamise ajaks on Helsingi ulikooli dateerimislaboris tehtud 8 dateeringut.

Dateerimiseks valiti savinoukillud kolmest tekstiilijaljendiga kaetud keraamikatuubist ning keraamikast, mis koostise ja pinnatootluse jargi liigitati algselt noorkeraamikaks (Kopu IA, joon 9). Viimane voeti asulakohast, kust on leitud ka (varast?) tekstiilkeraamikat, kuid selle pinnal ei olnud AMS-analuusiks piisavalt korbekihti. Pooled kaesolevas artiklis esitatavatest dateeringutest on keraamikast, mis parineb Ida-Eestist Peipsi jarve laanerannikul Emajoe suudmealal paiknevatest Akali ja Kullamae asualakohtadest (joon 1: 3-6). Need asulakohad on aga ka koige olulisemad ja rikkalikumad varase tekstiilkeraamika leiukohad kogu Eestis. Sealsete leidude pohjal see keraamikatuup uldse Lembit Jaanitsa poolt eristati ning planigraafia ja kaasleidude jargi dateeriti. Kaks savinoukildu parinevad asulakohtadest (Assaku Kukita ja Altkula, joon 7, 8), mille kontekst ja varasemad oletuslikud dateeringud andsid lootust saada teavet selle tekstiilijaljendiga keraamika "arengutest" nooremal pronksiajal ja eelrooma rauaajal.

Eestis esineb tekstiilijaljendeid neljal keraamikatuubil. Uksikjuhtudel on tekstiilijaljendid kantud hilise kammkeraamika noude pinnale, kusjuures neid on seni leitud vaid vahestest hilisneoliitilistest hilise kammkeraamikaga asulakohtadest. Erandina esineb tekstiilivajutisi ka muude parameetrite ja konteksti jargi noorkeraamikaks liigitatavatel savinoukildudel. Kahe savinoude tuubi puhul kasutatakse Eestis tekstiilkeraamika nimetust, jagades need varaseks tekstiilkeraamikaks ja tekstiilkeraamikaks. Esimene on dateeritud neoliitikumi lopuga (varasemas kirjanduses on varase tekstiilkeraamika algus ajaldatud 17.-16. sajandiga eKr) ja varase pronksiajaga, teine nooremast pronksiajast kuni eelrooma rauaaja keskpaigaga Pohja- ja Laane-Eestis ning keskmise rauaajaga Kagu-Eestis. Varane tekstiilkeraamika ja tekstiilkeraamika erinevad uksteisest nii vormimismassi koostiselt kui ka noude kujult ja ornamendilt.

Varane tekstiilkeraamika on valmistatud teokarbi-, kivipurru- voi taimse lisandiga segatud savist. Noud on suured ja samas suhteliselt vaikese lameda pohjaga. Kulgseinad on neil tugevasti profileeritud, servad kergelt valjapoole pooratud ja tavaliselt kulgseinast pisut paksemad. Noude valispinda katavad tekstiilijaljendid ja kammivajutised; ulaosas voib olla harvade ridadena lohke voi pulga umber keeratud nooriga tehtud vajutisi. Monikord on tekstiilijaljendit ka serval ja isegi nou sisepinnal.

Tekstiilkeraamika on valmistatud mineraalse lisandiga, peamiselt graniidigneissi ruhma kivimite purruga segatud savist. Potid on olnud erineva suurusega, erandina isegi ligi poolemeetrise suuava labimooduga. Noud on lihtsa kujuga: pustiste seinte ja vahe profileeritud servaosaga. Serv voib olla pooratud valja- voi sissepoole. Viimasel juhul voib uleminek kaelaosale olla rohutatud ka labi teravalt valjaulatuva nivendi. Tekstiilijaljend voib katta kogu valispinda, osa sellest voi olla vaid pohjal; monikord on tekstiilijaljend ka sisepinnal. Ornament (lohud, soor- ja napuvajutised, umber pulga massitud noori ning harva ka kammivajutised) on vahene ja tavaliselt vaid nou ulaosas.

Uued dateeringud Loona tekstiilijaljendiga kaetud hilisest kammkeraamikast ning Akali ja Kullamae varasest tekstiilkeraamikast osutavad, et need on olnud kasutuses uheaegselt hilisneoliitikumis (tabel 1). See kinnitab Jaanitsa poolt keraamika koostise ja leiukontekstide jargi tehtud oletust, et need tuubid on osaliselt samaaegsed ja saavad alguse neoliitikumi lopul. Nende kasutamise loppu ei voimalda saadud dateeringud maarata, kuid noorema pronksiaja muististes sellist keraamikat enam ei esine. See, et tekstiilkeraamika oma iseloomulike joontega oli valja kujunenud juba nooremal pronksiajal, sai selgeks Asva ja Iru kindlustatud asulakohtade uurimisel 1930. aastate teisel poolel. Altkula dateering tapsustab selle vanuse esialgselt veidi vanemaks kui 1000 aastat eKr.

Kuna esiajaloolised tekstiilid on sailinud vaid erandjuhtudel, on noori, longa ja riide voi riidelaadse materjali jaljendid keraamikal eriti oluliseks aluseks (kiviaja lopu ja metalliaja varasemal jargul Eesti alal seni isegi ainsaks voimaluseks), uurimaks tekstiilide valmistamisoskuse omandamist ja arengut. Jaljendite uurimine voimaldab teha otsuseid mitmete oluliste tekstiiltehniliste votete kasutuseletuleku aja, aga teatud maaral ka tehnilise taseme kohta.

Vaatlusaluses valimis maarati tekstiilijaljenditel koendi tuup, kasutatud longa labimoot ja voimaluse korral saie ehk keerd. Long voib olla kedratud paripaeva (S-keere) voi vastupaeva (Z-keere). Riidejaljenditel maarati koendi sidus ehk loime- ja koelongade ristumisskeem ning kanga tihedus, s.o loime- ja koelongade arv 1 cm pikkusel loigul (tabel 2). Jaljendid parinesid enamikus labasest (joon 10: a; 11: b-d), harvem ka ripssidusest riidest (joon 10: b, c; 11: a). Vahemalt uks tekstiilijaljend parineb noeltehnikas valmistatud esemelt (joon 12, 13). Varasemal keraamikal on tekstiilijaljendid tehtud ilmselt looduslikust kiudmaterjalist--nogese-voi niinekiust kootud riidega, hilisemal arvatavasti ka linasest voi isegi villasest materjalist riidega. Viimasele voib osutada noeltehnikana identifitseeritud tekstiilmaterjali jaljend Altkula asulakohast leitud keraamikal.

Dateeritud savinoukildudel esinevad tekstiilijaljendid on kantud noude pinnale erinevates tehnikates valmistatud riietega. Nendest enamiku moodustas labases koes materjal. Esindatud jaljendid olid nii uhesuguse voi lahedase tihedusega longasusteemidega kangast kui ka ripsist. Vaid ripsi puhul voib kindlamalt vaita kootud kanga kasutamist (koe- ja loimelongad on erineva labimooduga, mis noeltehnikas valmistatud materjali puhul on valistatud). Seevastu uhesuguse tihedusega longasusteemide ja samajameduste longadega tekstiilijaljendite puhul tekib probleeme: alati pole voimalik eristada kootud kangast noeltehnikas valmistatud materjalist (joon 13). Tanu tekstiilijaljenditega keraamika tapsustavatele AMS-dateeringutele nihkus taiuslikumate toovahendite--kedraga kasivartna ja kangaspuude--kasutuseletuleku aeg seni aktsepteeritust enam kui 1000 aastat varasemaks. Vahemalt neoliitilisele keraamikale on tekstiilijaljendid tehtud arvatavasti looduslikust kiudmaterjalist--nogese- voi niinekiust kootud riidega, hiljem kasutati ehk ka linast voi isegi villast riiet.

Kokkuvottes kinnitasid AMS-dateeringud konteksti, ennekoike Akali ja Kullamae planigraafia pohjal tehtud jareldust, et tekstiilijaljendeid hakati Eesti alal savinoudele tegema juba neoliitikumi lopul. Tosi, meie vanimad dateeringud--umbes 2700 aastat eKr--osutusid seni pakutust kull umbes 1000 aastat vanemaks. Tekstiilijaljenditega hiline kammkeraamika ja vormilt ning ornamendilt selgesti eristuv varane tekstiilkeraamika on laias laastus uheaegsed. Kuigi dateeringuid on veel vahe, osutavad need, et varast tekstiilkeraamikat on valmistatud kullalt pika aja jooksul.

Nooremasse, Eesti moistes tekstiilkeraamikasse toid uued dateeringud samuti korrektuure. Altkula keraamika dateering nihutab konkreetselt selle objekti savinoukillud ligi 1000 aasta vorra vanemaks varem esitatud oletusest, et tegemist voiks olla keraamikaga eelrooma rauaaja asulakohast. Samas osutab see koos Assaku Kukita dateeringuga, et paris noorema pronksiaja alguseks oli iseloomuliku vormimismassi koostise ja kujuga tekstiilkeraamika juba valja kujunenud.

Introduction

Throughout the history of archaeology, ceramic typologies have played an important role in compiling periodizations and chronologies of prehistory. So far the other methods of dating antiquities have not succeeded to replace ceramic typology though several of its aspects have been criticized. In reference to the settlement sites in the Baltic Sea region, ceramic typology is especially relevant to the investigation of the Late Neolithic and younger dwelling sites. These are, in many cases, no longer shore-related as were the settlement sites of previous periods. Due to the Post-glacial compensational land uplift, people rather often abandoned dwelling sites on the shore in order to move according to the regression or transgression of the water. (1) Many of the later settlement sites are, on the other hand, multiperiodical, having been inhabited throughout several prehistoric periods either continuously or discontinuously. Therefore, it is difficult to find any certain context for ceramics as well as for any other finds at these sites, and exact or, at least, more exact results are to be provided by typologies. However, typology as a method is inexact unless the types are related to calendar years obtained by scientific methods.

Nowadays, ceramic typologies have largely been corrected by the accelerator mass spectrometry (AMS) datings of the small amounts of charred organic remains (originating from carbonized food remains) preserved on the surfaces of potsherds, and by the calibration of the obtained dates into calendar years. The burnt organics and the clay vessel have been considered synchronous unless some extraordinary processes occurred in the ground after the deposition of the cultural layer.

In Sweden the AMS datings of the carbonized organics have been performed since the mid-1980s (Segerberg et al. 1991, 85), and in Finland since the beginning of the 1990s. Though in most cases merely single samples have been analysed, systematic AMS datings of the burnt organics on the ceramics in order to compile chronologies have been carried out as well, in North Finland for instance (Carpelan 2004). Dates of this kind become more and more numerous also in other regions, including the other parts of Finland (see e.g. Pesonen 1999; Lavento 2001a, b). However, only a few AMS datings have been made in the Baltic countries and Russia, which are very important areas also in reference to the Finnish ceramic types.

Considering the possibility of comparing clay vessels, one has to realize that the age identification of pottery is still based mostly on the general characteristics of the archaeological sites, find contexts of sherds and changes in the form and ornamentation of vessels. Problems also emerge because many typologies have been compiled decades ago. In the meantime, however, large amounts of new fords have been discovered. Therefore the typologies used in different countries no longer are unambiguously comparable. This is the case also with the Estonian Late Neolithic, Bronze Age and even Early Iron Age ceramics. New fords and contexts have come to contradiction with earlier typologies, necessitating their revision.

Starting-points of the ceramic dating project

The study of the Finnish, Karelian and Estonian textile-impressed ware has revealed that ceramics of the Sarsa-Tomitsa type on the east coast of the Baltic Sea, distinguished and dated to the Bronze Age more than half a century ago by Meinander (1954a, b), needs, in several respects, new specification. One of the diagnostic features of this pottery type, the textile impression, in fact occurs in several Neolithic ceramic groups. In the areas of present-day Estonia and Russia the use of textile-impressed clay vessels continued up to the middle of the Iron Age (Lavento 2001a). According to the existence of textile impression the sherds are considered to be of textile type. This type, therefore, involves several ceramic types regarded as separate groups at the present time (in Estonia Late Combed Ware, Corded Ware, Early Textile Ceramics, Textile Ceramics).

The "origin" of the textile-impressed ware is by no means less interesting: whether it originates from the tradition unambiguously related to some certain date and place, or is it rather a phenomenon independently "invented" in various regions of Europe. From the Finnish point of view, the suggestion of Meinander (1954b) that the ceramics of the Sarsa-Tonitsa type came from the south (from the Estonian area) as well as from the east (from the areas in the middle reaches of the Volga River in Russia), increased the relevance of the Estonian data.

The new datings also provide additional information about the sites where the dated potsherds come from. For half of the sites discussed in the present article, these dates are to be considered as the first dates obtained by scientific methods. Naturally, the dates are also important with reference to the history of textiles.

The main objective of dating the Estonian textile-impressed ware is to lay the foundations to a chronology based on the AMS dates of the textile-imprinted ware found in Estonia. The initial collection dated comprised 12 potsherds found at the oldest and most problematic settlement sites in Estonia. However, some samples taken from the charred organics of the potsherds did not contain enough carbon for dating and therefore we took additional new samples later. If the carbonized organics was not preserved on the textile-impressed sherds, the sample was taken, as an exception, from the ceramics of another type found at the same site. By the time of writing the current article, eight samples had been dated (Table 1) and, although the project has not yet come to an end, the results are interesting and worthy of immediate dissemination.

Dated ceramics and typologically relevant inferences

Finding places of the dated ceramics

For dating, we selected potsherds from among three different types of ceramics with textile impressions (the Late Combed Ware, Early Textile Ceramics and Textile Ceramics), and from the ceramics that, according to its consistence and surface treatment, was initially classified as the Corded Ware. The last type originates from the settlement site where the sherds of (Early?) Textile Ceramics are represented but provide no burnt particles sufficient for AMS dating. However, we expected to date in this way the find context of textile-impressed ware and find an answer to the question of whether the Estonian so-called Late Corded Ware is contemporaneous with the Early Textile Ceramics.

Half of the dates presented in the current article come from the pottery originating from the settlement sites of Akali and Kullamdgi in the boggy mouth areas of the River Emajogi on the west coast of Lake Peipsi, East Estonia (Fig. 1). In regard to the Early Textile Ceramics, these settlement sites are the most important and abundant in Estonia. On the basis of the finds from these sites, that ceramic type was first distinguished and, by means of horizontal stratigraphy and co-finds there, dated by Lembit Jaanits (AHHTC 1959, 140-149).

[FIGURE 1 OMITTED]

The potsherds, which, for Estonian archaeologists, represent the Textile Ceramics in the so-to-speak narrower sense, originate from settlement sites whose context and supposed dates seemed to be promising in revealing new information on the "development" of this pottery type in the Late Bronze Age and Pre-Roman Iron Age.

The Loona settlement site is situated in the western part of Saaremaa Island. Originally it was located on the shore of a small bay. The site was discovered in 1956 by Aita Kustin and was investigated in 1956 under the supervision of Kustin and in 1959 by Jaanits. The place had been inhabited in the Late Neolithic; the pottery is of the Late Combed Ware type, a small amount of it having textile impressions (Jaanits 1965, 30). The AMS datings of the bones of a seal and a pig date the settlement site of Loona to the average time span of 2900-2600 cal BC (2).

The Akali settlement site lies in East Estonia, on the bank of the River Akali, a tributary of the River Emajogi. The site was discovered in 1937 by Richard Indreko. Archaeological excavations were carried out there in 1938-1939 by Indreko and in 1949-1952 and 1968 by Jaanits. The cultural layer at the Akali settlement site covers a vast area of approximately 17 000 [m.sup.2], which, however, was not wholly in use at the same time. The settlement had been set up in the immediate vicinity of the river. As the level of the phreatic water rose, the place began to turn into a bog, and today the riverside part of the cultural layer is covered by a peat layer more than 2 m thick (Jaanits et al. 1982, 60). Therefore, in the course of time the inhabitants moved farther from the river. The oldest traces of life date to the Late Mesolithic but the site was nevertheless inhabited throughout the Neolithic as well as in the Bronze Age and at the beginning of the Iron Age (Jaanits et al. 1982, 43, 60). All the pottery types of those times are represented: the ceramics of the Narva type, Typical Combed Ware, Late Combed Ware, Corded Ware, Early Textile Ceramics, Textile Ceramics and other types. The only radiocarbon date (conventional) comes from the fire place where no ceramics was found, and it probably belongs to the Late Mesolithic period, the average of the dates being 5200 cal BC (3).

The Kullamagi settlement site is located in East Estonia, on the right bank of the River Emajogi, about 2 km west of the Akali site, on a sand elevation rising a bit higher than the surface of the surrounding marsh. The site was discovered in 1938 by Indreko. In 1951-1952, archaeological excavations were conducted by Jaanits. The cultural layer covers a vast area of about 10 000 [m.sup.2]. The place was used as a dwelling site from the Middle Neolithic. The Typical and Late Combed Ware, Early Textile Ceramics and other pottery types have been found there.

The Assaku Kukita settlement site is situated near Tallinn in North Estonia. The site was discovered in 1979 by an amateur archaeologist Oskar Raudmets. Two fire places were noted at the site, which was already damaged by land amelioration works, and the approximate area of the cultural layer was ascertained as 20-30 x 50 m (Lougas 1979). No archaeological excavations have been performed at the site.

The Altkula settlement site is situated in Southwest Estonia, on the high bank of the River Pdrnu. The small settlement site was discovered in 1972 by Vello L6ugas. A few potsherds, including these of the Textile Ceramics, were collected in the vicinity of a fire place that was destroyed by construction works (Jaanits et al. 1982, 176). No archaeological excavations have been carried out there.

The Kopu IA settlement site is situated in the western part of Hiiumaa Island. At the time of its establishment, it was located on the seashore. The site was discovered in 1981 by L6ugas and excavated in 1994, 1998 and 2000 by Aivar Kriiska. The place was inhabited in the Early Neolithic (ceramics of the Narva type) and in the Late Neolithic (the Corded Ware and textile-impressed ware, the specified type of the latter is not identifiable because of too small sizes of the sherds--Kriiska 2001). The radiocarbon dates (conventional) of charcoal collected in the hearths indicate only the early habitation phase, that is, 4500-4200 cal BC (4).

Characteristics of the ceramics and obtained dates

1. Al 4210: 649 Loona settlement site (Fig. 2).

Inclusions of the modelling paste: Shell debris and vegetable mixture (on the surfaces (5) and fractures, long impressions of fibres are observable).

Modelling technique: Modelled of bands (broad bands with N-type attachment (6)).

[FIGURE 2 OMITTED]

Shape and size of the vessel: Probably a pot having a rounded bottom; diameter of the rim approximately 44 cm; height about 30-40 cm; the rim is thinning, unprofiled; thickness of the walls 11-13 mm; thickness of the rim 8-9 mm.

Surface treatment and ornamentation: The interior is striated, without ornamentation; the exterior is textile-impressed to the full extent, the rim bears diagonal grooves.

Textile impression: The impressions have been made with fabric woven in repp technique (Fig. l la, Table 2). Both in the warp and the weft Z-spun yarns have been used, with the diameter of 1.5-2 and 3-3.5 mm, respectively. Thickness of the warp yarns has not been uniform. Weft yarns have been loosely spun but they are more uniform. In some places, the yarn has been flat and thus longitudinal unspun fibres are observable. The warp yarns were possibly made of nettle, and the weft yarns of bass (lime?).

Sample information: The sample was taken from the carbonized organics on the interior surface.

Date: 4165 [+ or -] 90 BP (Hela-751).

2. AI 4013: 8521 Akali settlement site (Fig. 3).

Inclusions of the modelling paste: Vegetable mixture.

Modelling technique: Modelled of bands.

Shape and size of the vessel: A pot; the rim is thickening, unprofiled; thickness of the walls 8 mm; thickness of the rim 11-12 mm.

[FIGURE 3 OMITTED]

Surface treatment and ornamentation: The interior is smooth, without ornamentation; the exterior is textile-impressed to the full extent, two lines of pits occur below the rim.

Textile impression: The impressions have been made with fabric woven in repp technique that had an S-spun warp 2-2.5 mm in diameter, and a weft 4-5 mm in diameter (Table 2). The thickness of the warp yarns has not been uniform. The weft yarns have been loosely spun but they are uniform. In some places the yarn has been flat, unspun segments with longitudinal fibres are observable. The warp was probably made of nettle, but the weft material bass (lime?) could have been used as well.

Sample information: The sample was taken from the carbonized organics on the interior surface.

Date: 4055 [+ or -] 40 BP (Hela-752).

3. AI 4045: 1052 Kullamagi settlement site (Fig. 4).

Inclusions of the modelling paste: Vegetable mixture.

Modelling technique: Modelled of bands?

Shape and size of the vessel: A pot; the rim is thickening and curved outwards; thickness of the walls 8 mm; thickness of the rim 8.5-11 mm.

Surface treatment and ornamentation: The interior is smooth, without ornamentation; on the exterior a zigzag ornamentation of comb impressions (8 zones of zigzags) occurs on the rim, a textile impression is found below the rim, on the side wall.

[FIGURE 4 OMITTED]

Textile impression: The impression is insufficiently preserved to be precisely identified.

Sample information: The sample was taken from the carbonized organics on the interior surface.

Date: 4140 [+ or -] 70 BP (Hela-754).

4. AI 4045: 1109 Kullamligi settlement site (Fig. 5).

Inclusions of the modelling paste: Vegetable mixture.

Modelling technique: Modelled of bands, 2-3 mm in width.

Shape and size of the vessel: A pot having a flat bottom; the rim is thickening and curved outwards; diameter of the orifice is approximately 40 cm; thickness of the wall 7-9 mm; thickness of the rim 14 mm.

Surface treatment and ornamentation: The interior is smoothed 7, on the rim there is a zigzag line of comb impressions; the exterior sides are textile-impressed to the full extent, the rim has horizontal and zigzag lines made by comb stamp (three single zigzag lines and four double horizontal lines).

Textile impression: The impression has been produced by textile made in tabby weave (?) (Fig. 11c, Table 2). The fabric has been quite dense; the thread count both in the warp and the weft is 6-8 per 1 cm, which is the largest number among the examined impressions. The warp was likely made of Z-spun yarn, while in case of weft yarn the spun was unidentifiable. The yarn has been tightly spun and is uniform in thickness. If the fabric has been woven on a loom (supposedly on the upright loom), it would be a firm evidence of fully established high-grade weaving skills and advanced technical implements. However, technically it is still possible that a material made in a simpler needle-netting technique was used as the matrix of the impression, since the impression of that material is very similar to the imprint of the fabric made in tabby weave (Fig. 13).

[FIGURE 5 OMITTED]

Sample information: The sample was taken from the interior surface of the vessel.

Date: 3605 [+ or -] 40 BP (Hela-755).

5. AI 4013: 3061 Akali settlement site (Fig. 6).

Inclusions of the modelling paste: Vegetable mixture.

Modelling technique: Modelled of bands?

Shape and size of the vessel: A pot having a flat bottom, the latter with salient edge; diameter of the bottom 10 cm; thickness of the walls 7-15 mm.

Surface treatment and ornamentation: The interior is striated, no ornamentation; the exterior is textile-impressed to the full extent, the salient bottom edge carries two lines of pits.

Textile impression: The impression is poorly examinable. Yarn 2.5-3.5 mm in diameter has been used as the warp, and yarn 3.5-4 mm in diameter as the weft (Table 2). The thickness of the warp yarns has not been uniform; the spun is unidentifiable (Fig. 11b).

[FIGURE 6 OMITTED]

Sample information: The sample was taken from the carbonized organics on the interior surface.

Date: 4155 [+ or -] 65 BP (Hela-761).

6. AI 5030: 1-2 Assaku Kukita settlement site (Fig. 7).

Inclusions of the modelling paste: Rock debris.

Modelling technique: Modelled of bands (broad bands with U-type attachment).

Shape and size of the vessel: A pot; the rim is slightly curved outwards; diameter of the rim about 37-40 cm; thickness of the walls 11-12 mm; thickness of the rim 12-13 mm.

Surface treatment and ornamentation: The interior is smooth; the exterior is textile-impressed to the full extent, pits occur on the neck.

Textile impression: A fabric in tabby weave has been used as textile matrix (Fig. 11d, Table 2). The diameter of the unevenly spun yarn has been 2-2.5 mm in the warp and 2-3 mm in the weft. The density of both thread systems has been similar: in the warp 6-8 threads and in the weft 4-6 threads per 1 cm.

Sample information: The sample was taken from the carbonized organics on the interior surface.

Date: 2765 [+ or -] 50 BP (Hela-837).

[FIGURE 7 OMITTED]

7. AI 4592: 1 Altkula settlement site (Fig. 8).

Inclusions of the modelling paste: Rock debris.

Modelling technique: Modelled of bands?

Shape and size of the vessel: A pot; the rim is thinning and slightly curved outwards; thickness of the walls 7-9 mm.

Surface treatment and ornamentation: The interior is smooth, without ornamentation; the exterior is textile-impressed to the full extent.

Textile impression: The material used for making the impressions has probably been made in the needle-netting technique (Fig. 12b, Table 2). The yarn, tight and Z-spun, was 2-2.5 mm in diameter.

Sample information: The sample was taken from the outer surface.

Date: 2885 [+ or -] 45 BP (Hela-838).

[FIGURE 8 OMITTED]

8. AI 6007: 1734 Kdpu IA settlement site (Fig. 9).

Inclusions of the modelling paste: Vegetable mixture.

Modelling technique: Modelled of bands (having U-type attachment).

Shape and size of the vessel: A pot; diameter of the rim approximately 20 cm; thickness of the walls 1.2 mm.

Surface treatment and ornamentation: The interior is smooth, without ornamentation; the exterior is smooth, without ornamentation.

Textile impression: No impression.

Sample information: The sample was taken from the interior surface.

Date: 5540 [+ or -] 55 BP (Hela-843).

[FIGURE 9 OMITTED]

New dates and the ceramic typology

In Estonia, textile impressions occur on four types of ceramics. On rare occasions, textile impressions can be found on the surfaces of Late Combed Ware vessels. However, so far only few Late Neolithic (i.e. later than 3000 cal BC in date) settlement sites containing the Late Combed Ware pottery have yielded such fords. Sherds of this kind could be classified as the Late Combed Ware according to the composition of their modelling paste and modelling technique, as well as according to the shape and ornamentation (or the lack of the latter) of the vessel. In the frames of the current project, one potsherd supposedly of the Late Combed Ware type, found at the settlement site of Loona, has been dated.

In exceptional cases, textile impressions occur on the clay vessels which, on the basis of the other parameters and context, could be classified as the Corded Ware. In addition to the observations made in Estonia (Kriiska 2000, 66), the same phenomenon has been noted in Finland (Edgren 1970, 33) and in Latvia (BaHKHHa 1980, 56) as well. Also the Late Neolithic Polja Ware in Finland bears occasionally textile impressions (Meinander 1954a, 165-166).

In Estonia, the term "textile-impressed ware" is used in reference to two types of ceramics: (1) the Early Textile Ceramics and (2) the Textile Ceramics (8). The former has been dated to the end of the Neolithic (in previous publications, the introduction of the textile-impressed ware has been dated to the 17th-16th centuries BC--[TEXT NOT REPRODUCIBLE IN ASCII.] 1959, 301) and to the Early Bronze Age, and the latter to the time span from the Late Bronze Age up to the middle of the Pre-Roman Iron Age in North and West Estonia (Valter Lang pers. comm. 15.03.2005), and up to the Middle Iron Age in Southeast Estonia (until the 6th century AD--[TEXT NOT REPRODUCIBLE IN ASCII.] 1997, 402). The Early Textile Ceramics and the Textile Ceramics differ from each other in the composition of the modelling paste as well as in the shape and ornamentation of the vessels.

The Early Textile Ceramics items are made of clay mixed with shell or rock debris, or with vegetable admixture. The vessels are large, although they have relatively small and flat bottoms. Their walls are strongly profiled; the rims are slightly curved outwards and, normally, a little thicker than the side walls of the vessel. The exterior sides of the vessels are covered by textile imprints and comb impressions (mostly in straight lines or in zigzag); the upper part of a vessel may bear sparse lines of pits or impressions made by cord coiled around a stick. Sometimes textile impressions are found on the rims and even on the interior sides of the vessel ([TEXT NOT REPRODUCIBLE IN ASCII.] 1959, 143-148).

The Textile Ceramics vessels are made of clay tempered with mineral admixture consisting mainly of rock debris originating from the granite-gneiss group. The pots have been of various sizes, in exceptional cases even with the rim half a meter in diameter (Indreko 1939, 32). The shape of the vessels is simple: the walls are upright and the rims are slightly profiled. The rims are curved either outwards or inwards. In the latter case, the transition into the neck of the vessel is emphasized by a carinate extending sharply outwards (Lang 1991, 46). The textile impression covers either the entire exterior of the vessel or part of it, or is found only on the bottom. Sometimes textile impressions occur on the interior surface of the vessel as well. Ornamentation is relatively scanty and occurs usually only on the upper part of the vessel. The ornamentation elements include pits, circles, wound cord and finger-tip impressions, rarely also comb imprints (Vassar 1939, 80).

The new dates of the textile-impressed Late Combed Ware from Loona, and these of the Early Textile Ceramics from Akali and Kullamdgi, indicate that both pottery types have been in use simultaneously in the Late Neolithic. Thus, they confirm the supposition made by Jaanits on the basis of the composition and find contexts of the ceramics that these types are partially synchronous, and that they first appear at the end of the Neolithic (Jaanits 1955, 181). The achieved dates do not enable us to ascertain the end date of these pottery types but, anyhow, the Late Bronze Age sites no longer contain this kind of ceramics.

The data from the settlement sites located in the mouth area of the River Emaj6gi suggest that the Early Textile Ceramics and the Textile Ceramics have been "genetically" connected, i.e. merely the shape of vessels and composition of their modelling paste changed in the course of time. The fact that the Textile Ceramics in its characteristic features was fully formed already by the Late Bronze Age became evident by the investigations of the fortified settlements of Asva and Iru in the second half of the 1930s. The sample from Altkula provided a more exact date for the matter in question by locating this a little earlier than 1000 cal BC in the temporal scale.

Conclusions from material-technical analysis of textile impressions

Introductory remarks on the history of textiles

Concerning the oldest textile fabrics (in pure technical sense), references could be made to the fishing-nets made of bass or any other material, and to the other net-like braided artefacts that, evidentially, were in use in the Late Palaeolithic already. A find of the same kind from Estonia, the net remains and floats of pine bark found from the bog in Narva Siivertsi, is somewhat younger, dating to the Late Mesolithic (Indreko 1931). The find from Antrea Korpilahti in Karelian Isthmus (Palsi 1920), the net remains from Nidlose and Ordrup bogs in Denmark (Becker 1941, 131; Hald 1980, 127, fig. 118) and from some other places belong to the same period. Net remains found from the settlement sites of Sventoji in Lithuania (Rimantiene 1979, 73-78) and Sarnate in Latvia ([TEXT NOT REPRODUCIBLE IN ASCII.] 1970, 94-95) date to the Neolithic. The fishing-net of Antrea had been woven of common sallow (Salix cinera) bass (Kujala 1949), and these from Siivertsi, Sventoji and Sarnate of lime (Tilia cordata) bass (Indreko 1931, 56; BaHKHHa 1955, 144; 1970, 95). It is probable that already in those times, besides the fishing-nets also mats were braided, and perhaps some parts of the clothing as well. The Antrea net sheet wide of doubled bass yarn was approximately 27 m long and at least 1.3 m wide (Palsi 1920, 17). It is hardly possible to produce this amount of high-quality yarn without special tools and, therefore, the use of a spinhook or even of a spindle already in the second half of the 9th millennium cal BC ([sup.14]C dates--Takala 2004, 151) should be assumed.

Along with the invention and development of new technical methods, the importance of various textiles in the everyday life increased. Unlike other materials used in prehistoric times, unfortunately very few textiles have preserved up to now. The oldest textile finds in Estonia (fragments of woollen stuff and bands) originate from as late as the Roman Iron Age. Therefore, the imprints of strings, yarn and cloth or cloth-like materials (mats for instance) on the ceramics provide, as a matter of fact, the only opportunity to have insight into the history of mastering and developing textile manufacture in Estonia and the neighbouring areas at the end of the Stone Age and in the early Metal Age. The study of impressions yields conclusions about the time of the emergence of several important textile manufacturing techniques and, to a certain extent, about their technical level. The AMS dates of textile-impressed ceramics allow us to suggest that some more advanced tools, such as hand spindle and weaving loom, came into use more than 1000 years earlier than hitherto believed.

Methods and results of the study of textile impressions

As the textile impressions on potsherds are in negative form, the surface imprints of the impressions were taken in order to obtain the reverse image using the dental modelling wax "Astynax". The wax plates were softened by hot air. The gloss of the imprint was reduced and a light lustreless surface achieved with talc powder. A darker foil was achieved with extremely fine charcoal powder "Kindrus" used in photography. The image was examined in aslant falling light under binocular magnifier equipped with micro-measure. As in several cases more than just one sherd of a particular vessel were available, it was possible to make complementary analyses in order to check the initial results. The averages of the obtained results are presented in Table 2.

By examining the textile impressions the type of weave, diameter of the yarn and, if possible, strand or spun of the latter, were identified. Yarn could have been spun clockwise (S-spun) or counter-clockwise (Z-spun). The binding of the fabric, i.e. the crossing-scheme of the warp and weft threads, as well as the density of the cloth, i.e. the number of the warp and weft threads in the section 1 cm long was ascertained. The impressions mostly originated from the fabric in tabby weave (Figs. 10a; 11b, d). At least one textile impression has resulted from a material made in the looped needle-netting technique (Fig. 12b). Tabby is the simplest weave when the weft passes alternately over and under the warp (Fig. 10a). The warp and weft are often of the same thickness and the distance between the threads is equal. Repp is a variation of tabby; the analysis is the same but one thread system is set closer than the other, or the warp and weft threads are of different thickness (Fig. 10b, c). Both the impressions of tabby and repp weave were represented. Looped needle-netting or simply needlenetting is a kind of sewing based on loops or meshes combined in various ways (Figs. 12, 13).

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

[FIGURE 13 OMITTED]

Materials of the textile matrices

In the forest zone the oldest sources for fibre have been, in all probability, bass and nettle. As mentioned above, fishing-nets stranded of bass cords were in use in the Mesolithic already. Later on, the bass was often used for making ropes. For instance, remains of bass cords frequently occur in the cultural layers of medieval towns.

Probably at the same time also nettle (Urtica diocia) came into use as a source for fibre. Together with hemp and hop, nettle belongs to the nettle order (at the same time, hop belongs to the Cannabaceae family). Nettle and hop are the components of the few remains of the Estonian relic flood plain forests. As the flood plain forests were among the first landscape components in the Estonian vegetation, which became influenced by human impact, the question emerges whether these species are the relics of the first plants naturalized in our region (Laasimer 1965, 74).

The North Siberian peoples still used thread of nettle fibre for sewing as late as at the beginning of the 20th century. In Europe, the nettle was utilized as an additional fibre source in Germany at the time of World War I (Stokar 1938, 57). In Finland the fabric of nettle fibre woven on handloom was used as wrapping material at the time of the Winter War and the War of Continuation (Leena Tomantera pers. comm. 2002). Remains of the Neolithic and Bronze Age nettle fabric have been discovered in Denmark (fords from Slotshoj and Voldtofte). In North Europe, the nettle has been present since the Boreal climatic period (Tolonen 1981, 216; Hald 1980, 127); as a nitrofile, it grows especially willingly in the vicinity of human settlements. According to the description by U. T. Sirelius, transmitted by Manninen (1929, 305), in summertime the winter huts of the Khantys are "... often buried in a thick nettle coat that billows like a cornfield around them". Probably the picture was the same at the winter camps of hunterfisher-gatherers of that time in our region as well. It is unlikely that the fibre source so easily obtainable and growing en masse remained unexploited. In some places, the Khanty and Mansi peoples of West Siberia braided nets of nettle fibre and wove nettle fabric as late as at the outset of the 20th century. For that purpose, they collected nettles after moving to their winter huts in autumn; subsequently the plants were sheaved and set under the eaves to wither (Manninen 1929, 305). It seems that trampling in the places where the nettle was growing en masse was wittingly avoided in order to protect these fibre plants.

To obtain fibre, the withered nettles were retted and barked with the help of a small wooden artefact, ethnographically called luda, or teeth (Manninen 1929, 306). Subsequently the material extracted from the pith was pounded with a pestle and scutched using wooden or bone knives (Hald 1980, 125). Also in the Far East and North America, fishing-nets were braided and cloth was woven of yarn spun of nettle fibre. Several languages contain evidence of the exploitation of the nettle as a fibre plant. Thus, once the original meaning of Finnish pellava (flax) was "the nettle" (Toikonen et al. 1962, 514). In Latvian, an analogous relation exists between nettle--natre, and linen--natns (Muhlenbach 1925, 702).

The hop (Humulus lupus), too, could be considered as an important natural fibre plant. In Estonia, the hop started to spread in the Atlantic climatic period. Probably rather soon and along with the increasing cultivation of barley, hops became known as an appropriate admixture in making beer. However, direct evidence of the exploitation of the hop as a fibre source is absent in the archaeological record as well as in written sources and folklore. Yet, the hempen fibre is long and, due to various vegetable poison substances, it is resistant to mildew, especially to that caused by moisture (the observations made by Juri Peets). If it was still used as a fibre source, it was presumably processed in the same manner as was flax or nettle. The other researchers, too, note the exploitation of the hop as a fibre plant (e.g. Hald 1980, 130).

The flax (Linum usitatissimum) is deservedly looked upon as one of the oldest cultured plants in the world. In Europe the evidence of its use has been obtained, for instance, from the Neolithic pile-dwellings in Switzerland and from the ancient settlement sites of the same age in East and Central Europe, in Belorussia and Germany, respectively (TEXT NOT REPRODUCIBLE IN ASCII.] 1969, 87; La Baume 1955). However, no analogous data are available from Estonia so far. The oldest remains of linen cloth from Estonia, small fragments of a fine-woven fabric, were found along with the hoard of Pilistvere, dating to the 6th century AD (Moora 1957, 203). In the lake sediments of South Finland (Hame) and North Sweden, the flax pollen appears relatively late as well, not until the 5th century AD. The same is also valid for hemp (Canabis sativa). Somewhat earlier, in the Pre-Roman Iron Age at the latest, flax cultivation had started in the Netherlands and North Germany (Lempidinen 2003, 330).

In general, all the researchers who have studied textile-impressed ware have unanimously agreed that as the textile matrix, the fabric woven of plant fibre was used. First of all linen or hempen cloth has been considered, but the use of the nettle has been admitted, too (e.g. Laul 1966, 99). Woollen fabric, on the other hand, was a priori considered to be too soft for obtaining a clear imprint. (9) And yet, the impression on the ceramics found from the Altkula settlement site, made with the fabric that was identified with a considerable certainty as having been made in needle-netting technique, possibly indicates the use of woollen cloth. Whereas needle-netting has been often used for producing things of woollen yarn (mittens, socks, caps, etc.), J. Peets took, as an experiment, some wax imprints from modern woollen mittens made by needle-netting and knitting, and from a rather coarse medieval cloth fragment. While experimenting on dry textiles, difficulties arose in removing the hardened wax. Wet fabric, on the contrary, detached itself from the wax easily, leaving distinct and detailed weave imprints on it. Therefore, it is possible that the woollen fabric was applied as a textile matrix in the prehistoric times, too. From the technological point of view, it is possible that the textile impressions were pressed onto the surface of the clay vessel with a mittened hand. The mitten could have been sewn of woven fabric or made by needle-netting.

The relatively late start of flax cultivation in Estonia (presumably not much earlier than in Finland) excludes the possibility that the textile impressions on the Neolithic and Bronze Age (10) clay vessels originate from the fabric made of flax or hemp fibre. The impressions on the earlier textile-impressed ware have probably been made with the fabric of natural fibre material (of nettle or bass fibre), and these on the Textile Ceramics presumably also with linen or woollen cloth.

Conclusions

The new dates obtained confirmed the conclusion made on the basis of find context, first of all on the grounds of the horizontal stratigraphy of the Akali and Kullamdgi sites, that in the Estonian area textile impressions were made on clay vessels already at the end of the Neolithic. However, the earliest of the recent dates--2800-2700 cal BC--turned out to be approximately 1000 years older than hitherto assumed. The Late Combed Ware involving textile impressions and the Early Textile Ceramics that are clearly distinct from each other both in shape and ornamentation are, in general lines, contemporaneous. Although the dates are few in number, they still indicate that the Early Textile Ceramics has been produced during a considerably long period of time. For the present, the temporal distance between the earliest and the youngest date is a little less than 1000 years. Let us mention here that the textile-impressed sherds of the Corded Ware found in the Riigikula XIV settlement site, which are the only sherds of this kind radiocarbon dated on the basis of charcoal collected from the site, originate from the same period (about 2500 cal BC in date (11)).

The new dates made some corrections also in regard to the younger pottery type or, as common in the Estonian research tradition, to the Textile Ceramics. The sample from Altkula shifts the date of the potsherds from this particular site approximately 1000 years back in time, as up to now it was assumed that the sherds originated from the Pre-Roman Iron Age settlement site (Jaanits et al. 1982, 176). At the same time this date, as that from the Assaku Kukita site, indicates that the Textile Ceramics with its typical form and composition of clay mass was fully established already by the very beginning of the Late Bronze Age.

The potsherd found from the Kopu IA settlement site was dated to the Early Neolithic, indicating that the fragment was not of the Corded Ware but of Narva type. This confirms once again that it is easy to be mistaken while identifying ceramics of similar composition, modelling technique and surface treatment, without additional support from the differences in typical ornamentation.

The textile impressions on the dated potsherds seem to have been pressed onto the surfaces of the vessels using fabrics made in different techniques. The majority of the impressions were made with fabric in tabby weave. The imprints studied were made with fabrics that had the same or similar density of the thread systems, as well as with repp. Only in case of repp it is possible to assert with 100% certainty that the fabric has been produced by weaving (the weft and warp yarns are different in diameter, which would be excluded in case of needle-netting technique). But problems emerge with textile impressions made with fabrics having the same density of the thread systems and threads of the same thickness, since it is not always possible to distinguish the woven fabric from the needlenetted material.

At least the textile impressions on the Neolithic pottery were made using fabric woven of natural fibre material, that is, of nettle or bass fibre. Later, the linen or even woollen cloth could have been used as well.

Acknowledgements

The radiocarbon dating was performed in the laboratory of Helsinki University under the conduct of Professor H6gne Jungner. The analyses were financed by the Jenny and Antti Wihur Foundation. Funding for this research was provided by grant No. 4558 of the Estonian Science Foundation. Thanks are extended to Kristel Kulljastinen for preparing the illustrations.

References

Becker, C. J. 1941. Fund of Ruser fra Damnarks Stenalder.--Aarboger for Nordisk Oldkyndighed og Historie, 1, 131-148.

Bronk Ramsey, C. 2005. OxCal (computer program). Version 3.10. The Manual (available at http://www.rlaha.ox.ac.uk/oxcal/oxcal.htm).

Carpelan, C. 2004. Environment, archaeology and radiocarbon dates.--Iskos, 13, 17-45.

Edgren, T. 1970. Studier Ever den sn6rkeramiska kulturens keramik i Finland. (SMYA, 72.) Helsingfors.

Europaeus-Ayrapaa, A. 1930. Die relative Chronologie der steinzeitlichen Keramik in Finnland I.--Acts Archaeologica, 1, 165-190, 205-220.

Hald, M. 1980. Ancient Danish Textiles from Bogs and Burials. A Comparative Study of Costume and Iron Age Textiles. Copenhagen.

Indreko, R 1931. Kiviaja vorgujaanuste leid Narvas.--Eesti Rahva Muuseumi aastaraamat, VII. Tartu, 48-67.

Indreko, R 1939. Asva linnus-asula.--MEL, 17-52.

Jaanits, L. 1955. Neoliitilised asulad Eesti NSV territooriumil.--MAL, 176-201.

Jaanits, L. 1965. Uber die Ergebnisse der Steinzeitforschung in Sowjetestland.--Finskt Museum, LXXII, 5-46.

Jaanits, L., Laul, S., Lougas, V. & Tonisson, E. 1982. Eesti esiajalugu. Tallinn.

Jaanusson, H. 1981. Hallunda. A Study of Pottery from a Late Bronze Age Settlement in Central Sweden. Stockholm.

Jaanusson, H. 1988. Beziehungen zwischen den Lausitzer and Nordischen Kulturprovinzen wdhrend der jungeren Bronzezeit.--Forschungen zur Problematik der Lausitzer Kultur. Warszawa, 171-177.

Kriiska, A. 1996. The Neolithic pottery manufacturing technique of the lower course of the Narva River.--Coastal Estonia. Recent Advances in Environmental and Cultural History. (PACT, 51.) Rixensart, 373-384.

Kriiska, A. 2000. Corded Ware Culture sites in North-Eastern Estonia.--De temporibus antiquissimis ad honorem Lembit Jaanits. (Muinasaja teadus, 8.) Tallinn, 59-79.

Kriiska, A. 2001. Stone Age Settlement and Economic Processes in the Estonian Coastal Area and Islands. Dissertation. Helsinki. httv://ethesis.helsinki.fi/iulkaisut/kultt/vk/kiiiska/

Kujala, V. 1949. Antrean Korpilahten kivikautisen verkon kuituaines.--Suomen Museo, 1947-1948, LVI.

Laasimer, L. 1965. Eesti NSV taimkate. Tallinn.

La Baume, W. 1931. Die vorgeschichtliche Handspindel and ihr Gebrauch. (Mannus, Erganzungsband, VIII.)

Lang, V. 1991. Uhe savinoutuubi ajaloost Loode-Eestis.--Artiklite kogumik. (Muinasaja teadus, 1.) Tallinn, 45-65.

Laul, S. 1966. Tekstiilijalgedest keraamikakildudel Eestis.--Pronksiajast varase feodalismini. Tallinn, 96-101.

Lavento, M. 2000. Some viewpoints on Early Textile Ceramics in the Baltic countries, Russia and Finland.--De temporibus antiquissimis ad honorem Lembit Jaanits. (Muinasaja teadus, 8.) Tallinn, 103-131.

Lavento, M. 2001a. Textile Ceramics in Finland and on the Karelian Isthmus. Nine Variations and Fugue on a Theme of C. F. Meinander. (SMYA, 109.) Helsinki.

Lavento, M. 2001b. Textile ceramics in Finland--recent perspective.--Acta Archaeologica, 1999, 59-78.

Lempiainen, T. 2003. Kasviarkeologiaa Auranjoen rannoila.--Kaupunkia pintaa syvemmalta. Turku, 323-340.

Lougas, V. 1979. Assaku Kukita asulakoht. (Manuscript at the Institute of History, Tallinn.)

Manninen, I. 1929. Soome sugu rahvaste etnograafia. Tartu.

Manninen, I. 1933. Die Sachkultur Estlands, II. (Opetatud Eesti Seltsi erivaljaanne, II.) Tartu.

Meinander, C. F. 1954a. Die Kiukaiskultur. (SMYA, 53.) Helsinki.

Meinander, C. F. 1954b. Die Bronzezeit Finnlands. (SMYA, 54.) Helsinki.

Moora, H. 1957. Varasemaid andmeid ketramisest ja kudumisest.--Eesti rahvaroivad XIX sajandist ja XX sajandi algult. Tallinn, 203-209.

Muhlenbach, K. 1925. Lettisch-deutsches Worterbuch. Riga.

Pesonen, P. 1999. Radiocarbon dating of birch bark pitches in Typical Comb Ware in Finland.--Dig it All. Papers Dedicated to Ari Siiriainen. Helsinki, 191-199.

Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W. et al. 2004. IntCal04 terrestrial radiocarbon age calibration, 0-26 cal kyr BP.--Radiocarbon, 46: 3, 1029-1058.

Rimantiene, R 1979. Sventoji. Narvos kulturos gyvenvietes. Vilnius.

Palsi, S. 1920. Ein steinzeitliche Moorfund bei Korpilahti Ksp. Antrea.--SMYA, XXVIII. Helsinki, 3-19.

Segerberg, A., Possnert, G., Arrhenius, B. & Liden, K. 1991. Ceramic chronology in view of [sup.14]C datings.--Laborativ Arkeologi, 5, 83-91.

Suriainen, A. 1974. Studies Relating to Shore Displacement and Stone Age Chronology in Finland. (Finskt Museum, 1973.) Helsinki.

Stokar, W. v. 1938. Spinnen und Weben bei den Germanen. (Mannes-Bucherei, LIX)

Takata, H. 2004. The Ristola Site in Lahti and the Earliest Postglacial Settlement of South Finland. Jyvaskyla.

Toikonen, H., Itkonen, E. & Joki-Aulis, J. 1962. Suomen kielen etymologinen sanakirja, III. Helsinki.

Tolonen, M. 1981. An absolute and relative pollen analytic study on prehistoric agriculture in South Finland.--Annales Botanici Fennici, XVIII, 213-220.

Vassar, A. 1939. Iru Linnapara.--MEL, 53-100.

[TEXT NOT REPRODUCIBLE IN ASCII.] B. 1955. [TEXT NOT REPRODUCIBLE IN ASCII.] ([TEXT NOT REPRODUCIBLE IN ASCII.] CCP).--MAL, 138-152.

[TEXT NOT REPRODUCIBLE IN ASCII.] B. 1970. [TEXT NOT REPRODUCIBLE IN ASCII.]. [TEXT NOT REPRODUCIBLE IN ASCII.].

[TEXT NOT REPRODUCIBLE IN ASCII.] B. 1980. [TEXT NOT REPRODUCIBLE IN ASCII.]. [TEXT NOT REPRODUCIBLE IN ASCII.], 47-58.

[TEXT NOT REPRODUCIBLE IN ASCII.] C. 1997. [TEXT NOT REPRODUCIBLE IN ASCII.]--[TEXT NOT REPRODUCIBLE IN ASCII.]. [TEXT NOT REPRODUCIBLE IN ASCII.], [TEXT NOT REPRODUCIBLE IN ASCII.] 1919-1997. [TEXT NOT REPRODUCIBLE IN ASCII.], 402-409.

[TEXT NOT REPRODUCIBLE IN ASCII.]--[TEXT NOT REPRODUCIBLE IN ASCII.], 4, 70-87.

[TEXT NOT REPRODUCIBLE IN ASCII.]--[TEXT NOT REPRODUCIBLE IN ASCII.], 76-93.

[TEXT NOT REPRODUCIBLE IN ASCII.]

(1) In the Baltic Sea region shoreline displacements were used for chronological distinction of, for example, the Combed Ware in Finland (e.g. Europaeus-Ayrapaa 1930; Siiriainen 1974) and the Pitted Ware in Sweden (shortly presented e.g. in Segerberg et al. 1991, 83).

(2) Here and henceforth, all the calibrations are based on the following sources: atmospheric data from Reimer et al. (2004); OxCal 0.10 Bronk Ramsey (2005); cub r:5 sd:12 prob usp[chron]. The base dates: 4270 [+ or -] 75 (Ua-4824) and 4050 [+ or -] 80 (Ua-4825) [sup.14]C years.

(3) The base date: 6255 [+ or -] 100 (TA-103) [sup.14]C years.

(4) The base dates: 5698 [+ or -] 70 (Tln-1901), 5604 [+ or -] 52 (Tln-1873), 5575 [+ or -] 50 (Le-5452), 5464 [+ or -] 96 (Tln-1898), 5460 [+ or -] 100 (Ta-2686), 5370 [+ or -] 68 (Tln-1871), 5330 [+ or -] 90 (Ta-493) [sup.14]C years.

(5) It is possible that the numerous fibriform impressions on the inner surface result from supporting the body of the vessel with a grass wisp while making the textile impression.

(6) Technological parameters defined as in Kriiska (1996).

(7) The interior surface is angular, possibly due to supporting the vessel with a hand while making the textile impression.

(8) This is by no means a generally accepted designation. Thus, for instance, the Late Bronze Age coarse-grained pottery from the East European forest zone (including Estonia), which could also have textile impressions on the surfaces, has been termed as the ceramics of the Tapiola type and of the Asva type (Jaanusson 1981, 122; 1988, 173).

(9) However, according to Russian researcher 1. Tshemay ([TEXT NOT REPRODUCIBLE IN ASCII.] 1981, 84), the basis for the extensive spread of the textile-impressed pottery in the Dyakovo Culture area was created exactly by the outset of sheep rearing and, along with this, by the use of wool in cloth manufacturing in the third quarter of the 2nd millennium BC in the East European forest zone.

(10) The oblong bone and antler artefacts found from the cultural layer of the Late Bronze Age fortified settlement of Asva, interpreted as combs for scutching the flax, were formerly considered as the oldest sign of flax cultivation in the Estonian area (Jaanits et al. 1982, 144). However, ethnographical parallels allow us to consider them reaping tools, so-called weed sickles. Summer crop, especially lodged barley, was weeded together with roots, using an obtuse sickle or a fragment of it. On Saaremaa Island, for example, such weed sickles were still in use at the outset of the 20th century (Manninen 1933, 180-181).

(11) The base date: 3970 [+ or -] 100 (Ta-2680) [sup.14]C years.

Aivar Kriiska, Chair of Archaeology, Department of History, University of Tartu (Tartu Ulikooli ajaloo osakonna arheoloogia oppetool), Lossi 3, 51003 Tartu, Estonia; aivar.kriiska@ut.ee Mika Lavento, Institute for Cultural Research, University of Helsinki, Unioninkatu 38 F, 00014 Helsinki, Finland; mika.lavento@helsinki.fi

Juri Peets, Laboratory of Geoarchaeology and Ancient Technology, Institute of History (Ajaloo Instituudi geoarheoloogia ja muinastehnoloogia labor), Ruutli 6, 10130 Tallinn, Estonia; jyri.peets@mail.ee

Table 1. The AMS dates of the carbonized organics collected from the
surfaces of the ceramics

Tabel 1. Keraamika pinnalt kogutud korbekihi AMS-dateeringud

Site Region Lab. No. [sup.14]C years
(store No.)

Loona Saaremaa Hela-751 4165 [+ or -] 90

(AI 4210: 649) Island
Akali East Estonia Hela-752 4055 [+ or -] 40
(AI 4013: 8521)
Kullamagi East Estonia Hela-754 4140 [+ or -] 70
(AI 4045: 1052)
Kullamagi East Estonia Hela-755 3605 [+ or -] 40
(AI 4045: 1109)
Akali East Estonia Hela-761 4155 [+ or -] 65
(AI 4013: 3061)
Assaku Kukita North Hela-837 2765 [+ or -] 50
(AI 5030: 1-2) Estonia
Altkula Southwest Hela-838 2885 [+ or -] 45
(AI 4592: 1) Estonia
Kopu IA Hiiumaa Hela-843 5540 [+ or -] 55
(AI 6007: 1734) Island

Site Calibrated * Calibrated age with
(store No.) age with the the probability of
 probability of 68.2% (cal BC)
 95.4% (cal BC)

Loona 2920-2480 2880-2700
(AI 4210: 649)
Akali 2860-2470 2840-2490
(AI 4013: 8521)
Kullamagi 2900-2490 2870-2620
(AI 4045: 1052)
Kullamagi 2130-1870 2030-1910
(AI 4045: 1109)
Akali 2900-2570 2880-2630
(AI 4013: 3061)
Assaku Kukita 1020-800 960-830
(AI 5030: 1-2)
Altkula 1220-920 1190-990
(AI 4592: 1)
Kopu IA 4500-4260 4450-4340
(AI 6007: 1734)

* Atmospheric data from Reimer et al. 2004; OxCal v3.10 Bronk Ramsey
(2005); cub r:5 sd:12 prob usp[chron].

Table 2. Results of the technical analysis of textile impressions on
the ceramics; wa--warp; we--weft

Tabel 2. Keraamika pinnal olevate tekstiilijaljendite tehnilise
analuusi tulemused

Site Binding Density, Spun
(store No.) thread/cm (S, Z)

 wa we wa we

Loona repp 5-7 3.5-5 Z Z
(AI 4210: 649)
Akali repp 6-7 3-4 S flat
(AI 4013: 8521)
Akali tabby 5-7 5-6 ? ?
(AI 4013: 3061)
Kullamagi tabby (?) 6-(8) 6-(8) S (?) ?
(AI 4045: 1109)
Assaku Kukita tabby 6-7 4-6 S S
(AI 5030: 1-2)
Altkula needle- -- -- Z --
(AI 4592: 1) netting

Site Yarn diameter,
(store No.) mm

 wa we

Loona 1.5-2 3-3.5
(AI 4210: 649)
Akali 2-2.5 4-5
(AI 4013: 8521)
Akali 3-4 2-3
(AI 4013: 3061)
Kullamagi 2-3 3-4
(AI 4045: 1109)
Assaku Kukita 2-3 2-2.5
(AI 5030: 1-2)
Altkula 2-2.5 --
(AI 4592: 1)