rL-tliES

r"[-Il.il-:lt,e-

NTEJ.

Lrnk'-:

! F,rli'tragad ::ih, i L-,.

rn:l-.

L'P- Si'-::..---fll|g::i In::rr li

.!: F[T DI'nil: iL - --

_:_.. _ t.

t:l.a.ii il::

n,d,: r- .-- i

ARCHAEOLOGICAL FIELDWORK IN ESTO\I.\ 2013. 11.J_11:

E

ARC HAE OL O GIC AL MON ITOEINGAT MI]NGAAND SEPA STREETSIN THE OLD TOWN OF VILJANDI

ANDRES TUAT]RI ANiT /KAP.;T METSOJATartu tll,ikool, fialoo ia arh,eol,oogia instituut, Arlr'eoloogia osakond (Unt'uersity

of Tartu, Instttuie of History and. Arch,aeology, Departnxent of Arch'aeology)'

Jakobi 2, 57074 Tartu, Estont'a; dndres.tDauri@ut'ee

INTRODUCTIONMunga and Sepa streets lie within the territory of medieval Viljandi. Monitoring was

caried out in connection with the lenovation of water and sewerage pipelines (Fig' 1)'

Earthworks were monitored. in the area of approximately 375 m2; the depth of the

trench for pipelines was \p to 2.2 m. In addition, two archaeological trenches were

made (Tvauri & Metsoja 2013). The size of trench 1 at the western part of Munga

street was 7.5 m2 and the size of trench 2

at Sepa street was 10 m2. The occupationlayers on both Munga and Sepa streetswere approximately 1.5 m thick.

At the eastern Part of Munga street,in a section between SePa and Lossistreets, archaeological investigationshave been carried out before the instal-iation of water pipeline in 1989 ffa1k1990). In the courtyard of Munga 4, on thenorthern side of the street, building of astorage shed necessitated a small-scalemonitoring work in 2011, during whichno medieval layers were unearthed(Tvauri 2011). At Sepa street, archaeo-

logical monitoring has been carried outin the southern end of the street where atrench was dug in 2000 to install a pow-

er cable (Haak 2005, fig. 3). There wasno knowledge about the thickness andcharacter of the occupation layer of thewestern part of Munga street and mostpart of Sepa street before the currentrnonitoring.

0 100m

Ftg. 7. Trenchfor pi'peli'nes, documented duringthe monttoring works at Sepa and Mungastreets, an'd arch,aeologtcal trench'es 1 and 2

Jn 1. Vifondts Munga ja Sepatdnaual wlrcoloo'gt li.se j drelets ahs e kiiigus d,okumenteentudkaet:ts ja uurt'miskaeuandld 1-2'

Drowinq Joonis: And res Tuouri

wffi!ffiffi&i:'

E:i..

::::

i:

\___.:Housc Sepa 8/ /Sepa I hoone

I

housing remainshoonemiirid

foundation of Sepa 6 tSepa 6 uutudqmetut

Trench 2 /Kaeuand 2

0 10om

Fig. 2, Locatzon of the trench of ptpelines, housingremains dnd, trench 2 at Sepa street.

Jn 2 . Torusttku kaeuise, aDastatud, hoonemiiiil Ld,eja kaetsand 2 asukoht Sepa tiinaual.

Drawing t Joonis: Andres Tuauri

T-etch 2 at Sepa street. Vi,ew from the uest.-,cr:dei.:ae,-andile2Sepatdnaualldtnepoolt.

:, -i,:o'-:-. Trorrri

A

MEDIEVALSTREET PAVEMENTS

Throughout the entire Munga street,one could see layers of red Devoniansand on top of which in most places cob-blestone pavement had preserved thathad originally covered the entire streetwithin the width of the current street.The pavement appeared 0.5-1 m be-low the current street level. The layerconsisting of cobblestone pavement andits sandy bedding was in some placesonly 10 cm thick (eastern end of Mungastreet) while in other parts - in the west-ern end of Munga street and the middlepart of Sepa street - it reached 50 cmand had three different layers of sand iso-lated by layers of soil. Thus, probably inplaces where potholes occured, pavementstones had been removed, depressions re-filled and the pavement restored.

In 1989, during.the investigations atthe eastern part of Munga street, a stonecannonball fallen onto the cobblestonepavement and having penetrated it wasfound in the Devonian sand layer. Theball was most likely deposited in the year1560 when Moscovian troops besieged Vil-jandi (Valk 1990, 11-12). Therefore, it canbe assumed that this pavement, similar ofwhich can be seen in many places insidethe medieval town wall of Viljandi, datesfrom the Middle Ages (see Haak 2008).

Cobblestone pavement was also vis-ible in the southern and middle part of Sepastreet (Fig. 2). In trench 2 there were up tofive layers ofDevonian sand bedding ofear-Iier pavements under the layer of preservedpavement stones. Thus, the cobbles ofSepastreet have been removed, the surface re-filled and flattened and the pavement re-stored at least four times. Trench 2 revealedthat at least during the use of the last pave-ment layer there had been a drain runningalong the central axis ofthe street (Fig. B).

-r_: _:ffi

The red ..:rd,the pavemen: miffi

the filling la-"-=r: ha few sherds _: s,-lm

fill layer under"::mthe youngest x,:ii,flremarkabie ar_.d.Mbeaker decor.a:il rings (Fig. 4) i.;=:-* il

largestnumbrr,Jliuers. manufaci *-:[.the 1Sth - 1i::- .=min Tartu. A f'e-,,. ^ug

found in the rr::*Erthe suburban :i-*ffirand in the cas:-*r dand the hill f;:- :r I(Kull 2010). fa. :ustreet was thc ._-_.c

the medieval t,m1 6

The lar-er.. --rnldating from th; ,*ruFragments of ;:::s616th centulr-. -.::ydIn addition. ft'aa.amrobtained. Thu;. :r*

I

Livonian War.. :--:.sr i

beginning of th. - ad

REMAI}S OMNAt Munga strei: :m&brown organic-r--:: @ground were fr,;:r L U

any constructir-The onl1- s: - -,* u

of a stove and a y:,,r*r, d

in trench 1 (Fig : ffitioned medier-a,,,::ru[the natural grr::: [ [strate that at ier;l mC

that area, but a :r:rlruqIn the miii,. m

of stone buildin;= lnu(.rrgs I and r). -:.+ Nside of Sepa str-: ,rn

street,rvonianles COb-

d thatr streetstreet.m be-e layer:nt andplaces

NlungaLe west-middle50 cm

and iso-nbly insementions re-

tions ata stonerlestoneI it waser. The;he yearrged ViI-e. it can"milar of,.s insideli. dates003).also vis-t of Separe up toug of ear-re-"ervedr of Separface re-ment re-nevealeda,st pave-runninglFig.3).

ARCHAEOLOGICAL MONITORING AT MUNG.\ A\D SEP.\ STREETS I\ \-ILJ.\\DI

,n" ;1"";-"iiT3,l"l*T,-,':i";:ff'",: d# fBthe fiIling layers between sand beddings iffi.a few sherds of stoneware and, from the ffij$**.,-

iil,';ru:.11:",i*"'l;;il1 ::'l'tr":l ffi#ffiremarkable artefact. a fragment of glass W#FrfT;iFbeaker decoraled with polycrome paint- ffie - t'1f,,.3lngs (Flg. +) \YeI'e Iouno. jn estonra. r'tre * .1Iarsest number of fragments of such beak- -#EE#

il:,T,T:'il1il::1i#,::l';,1"'*m ffiin Tartu. A few fragments have also been GWffiSffi 0 i.:found in the medieval town of Haapsalu,the suburban area of medieval Tallinn Fi.g.4. Fragmentof arned.teualdecoratedglass

and in the castles of Viljandi and Otepdi beaker'fromtrench2 dt sepa street'

and the hill fort of vooremhe at Pikasilla Jn4' Keskaegsemaah'nsutegaklaaspeeb'tkild

(KuII 2010). The fragment found at sepa ,u*rrni{ilffcnarkeskosastkaetsandtst2'street was the first from the territory of PhotolFoto:AndresVindt

the medieval town of Viljandi.The layers immediately on top of the pavement Iayer contained abundant finds

dating from the second half of the 16th century or the beginning of the 17th centurl"Fragments of earthenware pots of Pskov origin, characteristic to the second half of the

16th century, sherds of Russian glazed greyware and an iron heelpiece were found.

In addition, fragments of glazed redware, g).azed stove tiles and window glass were

obtained. Thus, the pavement in question got buried either during the events of the

Livonian War, after Moscovian troops conquered Viljandi in 1560 or some time in the

beginning of the 17th century during the Polish-Swedish Wars.

REMAINS OF MEDIEVAL BUILDI.IVGSAt Munga street, below the pavement and bedding layers in the up to 60 cm thick dai'k

brown organic-rich medieval occupation layer, several logs and stakes rammed into theground were found. However, the logs were not bound with any joint and did not formany construction.

The only stone construction unearthed at Munga street was a granite stone base

of a stove and a row of stones (possibly a base for a wooden waII) connected to it. found

in trench 1 (Fig. 5). Below the stove-base and contruction attached to it, the afolemen-

tioned med.ieval occupation layer continued. The composition of the layer deposlted on

the natural ground and the stove-base and remains of logs within this la1'er demon-

strate that at least before the middle of the 14th century there was no street or load on

that area, but a courtyard and some building instead.In the middle part and on the western side of the Sepa street the fi'ont rralls

of stone buildings were unearthed, observable approximately in the length of 25 m(Figs 1 and 2). The profi1e of trench 2 reveals that the stone building on the westelnside of Sepa street was erected only after the first (i.e. lowest) street pavement was

\\DRES TVAURI ANd KART METSOJA

built. Out of five pavement layers, fourwere younger than the house, and wereIaid right against the wall.

As revealed during monitoringworks, the northern end of Sepa street,reaching 18 m from the southern edgeof Munga street, had been covered withhousing in the Medieval Period. Nopavement layers were discovered here.Instead, remains of a 50 cm thick brickwall made of 9-10 cm thick bricks run-ning along the southern border of Mungastreet were unearthed. Remains of a hy-pocaust with heat storage and a clay floor

\ were also found. Whether the latter tw<r

were related to the aforementioned brickwaII remained unclear.

PRE.PAVEMENT MEDIEUALOCCUPATION LAYER

Throughout the entire Munga street,there was a medieval occupation layerthat had deposited before the paved streetwas constructed. The layer was up to 60cm thick and consisted mainly of darkbrown soil, dampish and rich in organics,

containing small fragments of bricks, charcoal and animal bones. Pieces of differentrvooden dishes, and fragments of leather shoes, along with wood shavings, hazelnutsheIIs, and even hay had preserved in the layer. Artefacts can be dated from the middleof the 13th to the middle of the 14th century.

Earthworks revealed pottery production waste in the western end of Munga streetthat had deposited immediately on top of natural soil surface. Sherds of earthenware.bloken and melted during firing, stem from the mid-13th century pottery workshop onthe n'estern side of Pikk street, some 50 m south of the current find spot. The pottery

"',-ot'kshop has been excavated archaeologically, both remains of kilns and pits contain-

:nt ploduction waste have been investigated (Tvauri 2000).-\t Sepa street, the soil deposit below pavement layers was in average only around

-,' tm thick (in trench 2 the thickness was up to 50 cm) and contained considerablv-:ss hnds than the one at Munga street. Most numerous among the finds were sherdsi -:lthenu-ale manufactured in the nearby Pikk street pottery workshop. As the soil

'.', '.. -ir"j- at Sepa street, no artefacts of organic material were obtained from there.F:,;:a-,-ati:,ns revealed that south of trench 2, in front of Sepa 6 building there was no

..--r. .::- - :: lar-el deposited before the street was paved. An approximately 20 cm thicki' :r - :i.'-r"k s,:,:] found there laying on top of natural sandy loam consisted most likel-v, :r:--.-:-- ::rars eai'th and a few occasional potsherds.

Fig. 5. Trench 7 at the uestern part of Mungastreet. Vtew from the east.

Jn 5. Vaadekaeuandile 7 Mungatdnat:aLti,iine o s as zda p o olt.

Photo I Foto: Andres Tra.u.ri.

ARC:

Natural soii ::::":lutIn some places. tl:=:".

'rrq,ith no finds on t,: :i:ii'ards southeast: a: ::+t'rom the KronstaC: _,1.u,,r

the southern part ,. i*1

CONCLUSIO.\-'\'chaeological nrr :_: -::rbout the street n=:- :1

We learned tl-..: _J

-lth - 16th centul:=: I

rad been renovatei r=;,r*:rents, in use in -: -.lom the second ha_: :: ;:he pavement laler'. .

,,ther constructions.Amedieval ct:-

-Ireet was found a: -_:..

- ontemporary Sepa - .:.*r=r'al Period; hei'e e :r-.{iiscovered. In addi:, --- ::'Sepa street. A nr=:'_= -

-hick and poor in h::- l--rnstruction of the r .. , ,.r

}.cl;nowledgemen t; T-,E,tropean Regional D. ,,"r'i,lt

REI'ERENCESHaak, A. 2003. Keskae:.- ' 'Lrla

- -.-iljandi Muuseumi uo.,,..r.*,,*- -_91.--{aak, A. 2005. Viljand: __- :, , Lu

:: ..miselt arheoloogiliste ..-_,im'-::seumi toimerised. i. '. - r,i,'ull. K. 2010. 13. -14. . ,- : t,- .:a,speekrid Eestis. Bak:-_. ._::,,,q

.',! :nuscript in TAAK.tTi'auri, A. 2000. Pihkva. , :lil. ..rrrs 13. sajandiL. g.l..

ARCHAEOLOGICAL MONITORING AT MUNGA AND SEPA STREETS I.\ \-ILJ.\\DI

Natural soil throughout the investigated area was grayish or yellow sandl- 1oam.

I: some places, there was a layer of (natural?) dark soil, approximately 10 cm thick

', tth no finds on top of natural sandy loam. The natural ground surface descends to-

';aLds southeast: at the western end of Munga street the ground surface is at 83.60 m

:r -rrn the Kronstadt zero-Ievel, at the eastern end of Munga street at 82.40 m, and at

:--. southern part of Sepa street only at 82.15 m.

CONCLUSIONS----":haeological monitoring works at Munga and Sepa streets provided new information,'. -,ut the street network and buildings of medieval Viljandi.

We learned that Munga street in its current location had existed already in the

-=,h - 16th centuries. The street had been paved with cobblestones and pavements

: : 1 been renovated repeatedly. Throughout the entire street, the youngest of the pave-

r :nts, in use in 1560, had been preserved. An organic-rich occupation layer dating::' . m the second half of the 13th century or frorh the 14th century was revealed under::,= pavement layers of Munga street, containing the base of a stove and remains of

:i::L constructions.-\ medieval cobblestone pavement, similar to the one documented on Munga

r.r'tret was found at the southern and central part of Sepa street. The northern end ofr aremporary Sepa street, however, was covered with construction during the Medi-,,r, ..1 Period; here a brick waII, a hypocaust with a heat storage and a clay floor were,1 ,=:or-ered. In additon, remains of stone buildings were found from the western side

i: Sepa street. A medieval occupation layer below pavement layers was only 10-20 cm

iLt:: t-i &nd poor in finds. It appears that there were no houses at Sepa street before the

r' :,.truction of the paved. street probably some time in the 74th - 15th centuries.

.t .,:owled.gement: This research was supported by the Europeant Union through the

,r - ,pean Regional Deuelopment Fund. (Centre of Excellence in. Culture Theory (CECD)

.RETEEENCESlifriLrr rk. -{. 2003. Keskaegse Viljandi tiinavate vdrk.' :::di Muuseumi aastaraamat 2002. Viljandi,

,llt,, ,t. -f. 2005. Viljandi linna kujunemisest)' ::'.-r.lt arheoloogiliste allikate pohjal. Narvatrri ....-:mi toimetised, 5. Narva, 17-28.ffi.'r... K. 2010. 13.-14. sajandi emailmaalingutegat.,t , ,: =;kl.rd Eestis. Bakalaureusetdo. Tartu' ' .::,'iptinTUAK.)

''ltl ' i rri. -{. 2000. Pihkva pottsepad Viljandis ja''' '- .. -3. sajandil. trJA, 4, 1, 21-30.

Tvaurio A. 2011. Aruanne arheoloogilisestjdrelevalvest Viijandis Munga 4 krundil puukuuritulemiiiiri vundamendisiivendi kaevamise j uules30. mail 2011. aastal. Tattt. (Manuscript in l'LA'\Tvauri, A. & Metsoja, K. 2013. Aruannearheoloogilisest jdrelevalvest vee- jakanalisatsioonitrasside raj amise juures \iilj andisMunga, Sepa ja Pikal tdnaval 2013. aastal.(Manusoipt in MA.)Valk, H. 1990. Aruanne arheoloogili-'testkaevamistest Viljandis Munga tdnava veetrassil1989. a. Tallinn. (Manuscript in TL--AK.I

!u

$li

$rI

i

::

i-,L{:}GIC.\L FIELDWORK IN ESTONIA 2013, 113_118

-IRHE O L O O GI LIN E t enn n nv etv nI'ILJANDT yANALTNNAS MaNGA JA sgpArANevet.lndres Tu auri ja Kiirt Metsoja

-\'heoloogiline jdrelevalve Viljandi keskaegse Iinna a1al, Munga ja Sepa tdrnaval toimus seoses vee- ja ka-naLsatsioonitorustike renoveerimisega 0n 1). Jdrelevalvet tehti u BT5 m2 suurusel alal. Lisaks rajati kakskaevandit: Munga tdnava lddneosas paiknenud kaevandi 1 pindaia oli 7,5 m2 ja Sepa tdnaval paiknenudkaevandi 2 pindala o1i 10 m2. Kultuurkihi paksus oli enamasti umbes 1,5 m.

Vaadeldavas piirkonnas on varem suuremahulisi arheoloogilisi uuringuid tehtud Munga tAnavaidaosas. Enne 2013. a jArelevalvet puudusid andmed kultuurkihi paksuse ja koostise kohta Mungatanava lAdneosast ja enamiku Sepa tdnava alalt.

Kogu Munga tdnava alal oli jzilgitav punase liiva kihistus, mi1le peal oli enamasti sdilinud muna-kivisillutis, mis on katnud Munga tdnavat selle praeguses laiuses. Kihistus oli 10-50 cm paksune,koosnedes iihest kuni kolmest mullaga eraldatud liivakihist. Varasemate uuringute pohjal on teada,et see, linnamtiiiriga piiratud Viljandi a1al paljudes kohtades jAlgitav sillutis p2irineb keskajast. SiI-lutis oli alles ka Sepa tdnava l6una- ja keskosas. Sepa tAnava kaevandis on sillutise a1l jiilgitavad nelivarasema sillutise liivapadjandit. Vhhemalt k6ige viimases sillutises on piki Sepa t6nava keskosakulgenud sadeveerenn (jn 3). \

Sillutise alused liivakihid olid teiutiihjad. Vaid sillutiseliiva vahelistest tAitekihtidest leiti m6nedkivikeraamiliste kannude killud. TnhelepanuvAhrseimaks leiuks on viirvilise maalinguga klaaspeekrikild (n 4), mis oletatavasti phrineb Veneetsias valmistatud nSust ning on esimene ."du tiit pi t"idViljandi linnast. Siilutise pealne kiht sisaldas rohkesti leide 16. saj teisest poolest. Seega on .illrti.mattunud kas Vene-Liivi s6ja ajal pArast Viljandi vailutamist moskoviitide poolt 1b60. aastal voi 17.saj alguses Rootsi-Poola s6dade kiiigus.

Munga tAnaval paljastus tdnavasillutise aluses keskaegses kihis palke ja maasse rammitud vaiu.Leitud palgid polnud omavahel tapiga seotud ega moodustanud konstruktsioone. Ainus kivikonst-ruktsioon oli Munga 4 hoone ees kaevandis avastatud maakividest ahjualils ja sellega seostatavkivirida (jn 5), mis v6is olla osa puithoone vundamendist. Looduslikuie aluspinnasele ladlstunud kihikoostis ja selles leiduv ahjualus ja palgijiiiinused nhitavad, et enne 14. saj keskpaika ei olnud selle1kohal tdnav v6i tee, vaid hooviala ja hooned.

Sepa tanava keskosas avastati thnava lziiinekr.iljel kivihoonete esiseinte miiririd, mis olid jdlgitavadu 25 m pikkuselt (n 2). Kaevandi 2 profiili p6hjal otsustades on Sepa tdnava lddnekiilje k"Ju"gr"kivihoone v6i hooned rajatud pdrast Sepa tdnava sillutamist esimese munakivisillutisega.

Selgus, et keskajal on olnud hoonestatud ka Sepa tdrnava pohjaots alal, mis ulatub Munga tAnavaiounakiilje joonest 18 m l5una poole. Siin ei avastatud tdnavasillutise kihte, kiill aga tellis;iitri jee-nus Munga tdnava 15unaserva joonel, kerishilpokaustahju jdanus ning savip6rand.

Keskaegsete sillutiste alused kihistused olid Munga tdnaval kuni o0 cm paksused ja koosnesidpeamiselt pruunist orgaanikarohkest mullast, mis mh sisaldas vAikeste laudkausikeste tukke, treitudpuittaldrikute katkeid, nahkjalatsite katkeid, laaste, heina jm. Sellest kihist saadud leiumaterialir.dib dateerida ajavahemikku 13. saj keskpaigast kuni 14. saj keskpaigani.

N{ainimist vdiirivad Munga tAnava ldAneosas Iooduslikult maapinnalt Ieitud lihtkedrakeraamilistesar-in6ude valmistusjdhgid: pdletusel punsunud savindukillud. Need pdrinevad oma leiukohast viie-kumne meetri kaugusel Pika tdnava liizinekiiljel 13. saj keskpaigas tegutsenud keraamikatocikojast.

Sepa tdnaval oli sillutisekihtide all lasuv orgaanikarohke mullakiht keskmiselt vaid paarikiimnesentimeetri paksune (kaevandi 2 alal kuni 50 cm paksune) ja sisaldas mArgatavait vdhem leide kuillunga tdnaval. Arvukaimaks leiuks olid lihtkedrakeraamiliste pottide killud. Kaevandist 2 16unapool polnud enne tdnava sillutamist kuituurkihti iildse ladestunud.

Loodu-qlik aluspinnas on k5ikjal uuritud alal hallikat v6i kollakat tooni saviliiv. Kohati o1i selielsf,ilinud kummekonna sentimeetri paksune looduslik mullakiht. mis leide ei sisairlanud.

ARCHAEOIN THE C

MARTIIi N.Tartu Ulik,.":,i",.

of Tartu. Itr"sr:

Jakobi 2. 51t,:4

VILLU K-{.D.Tall,tnna t-,t,:Riiiitli 6, I &t i I

ELIS TIIDCTartu Linnc:t.

INTRODC'This paper intr.c,i ,yrwso called 'Neu- C::-:excavations in 2_ -

water and se\^-erar:

HISTORICThe parish of Kos=quering North E=:: mrr

baptising of the :-r:*(Raam 1990, 119r ur:stone church. -\c."-r"riiseveral construcrr- - -erected in the 13i -,;(Fig 1: 2). At firsr ::tioned as a chanrn-the western gabtrtr. -1into the old eastey:limestone, the cha-:*:probably constlur:.1

The church r'-:iably the vaults of i:*was demolished a:-:nected with the na-,'"and the church br:decoration and a sr

1 In this article ::,- : r;umqlTiidu, and the r : -,--. ,,. mt