Roman Ru al Si e a S . Pe kina Ni a, No h Macedonia.
A chaeological and Geophysical In es iga ion
Pe a Tušlo á– Ma ek Ve čík– Pe o A djanlie – Michal Vágne – Jan Ha elka–
Dá id Šálka
ABSTRACT
Acombina ion o su ace su ey, geophysical p ospec ion and exca a ion was used o in es iga e asi e da ed
o he Roman pe iod in No h Macedonia nea by he Oh id Lake. The si e is placed in he e i o y o Leskoec
illage, in an a ea locally known as S . Pe kina Ni a. The exca a ion unco e ed s one ounda ions o ahouse,
wi h s one des uc ion inside he house and likely apa emen on i s ou e side. A he g ound le el inside
he house, jus abo e he bed ock in o which he ounda ions we e dug, agmen s o se e al s o age essels
(dolia) we e ound, some in si u, some in agmen s, sca e ed o e he inne space. The dolia, as well as he
accompanied ma e ial including a ious po e y classes, glass agmen s, bone pin and ew me als, migh be
da ed om he 1s ill 4 h c. AD. The same ch onology migh be an icipa ed o he in es iga ed house, which
e ealed only one cons uc ion phase. Due o he lack o diagnos ic ma e ial he cons uc ion o habi a ion
pe iod o he house canno be na owed down.
KEYWORDS
No h Macedonia; Oh id Lake; Roman pe iod; u al; se lemen ; geophysics; exca a ion.
INTRODUCTION
Since 2017 he Ins i u e o Classical A chaeology a Cha les Uni e si y in P ague in coope a ion
wi h he A chaeological Museum o he Republic o No h Macedonia has been conduc ing he
F on ie S udies p ojec o in es iga e p o o -his o ic and ancien habi a ion de elopmen s in
he egion a ound Lake Oh id in No h Macedonia. The esea ch wo k ini ially ocused on he
explo a ion and documen a ion o a chaeological and epig aphic indings beyond he u ban
cen e o Oh id, he ancien Lychnidos, which ha e so a a ac ed ela i ely scan esea ch
a en ion (Ve číke al. 2019). In addi ion, he majo i y o he si es eco ded in he local sec ion
o he gaze ee o a chaeological si es o No h Macedonia Археолошка карта на република
Македониja we e iden i ied based on s ay inds (AKM 1996). Sys ema ic exca a ions, which
could p o ide in o ma ion on he opog aphical loca ion, geomo phological se ings, ex en ,
in e nal s uc u e, and ch onology, ha e only a ely been conduc ed. As a esul , he e a e
se e al spa ial and ch onological in e ludes o he diach onic econs uc ion o se lemen
de elopmen in he Oh id egion (Ve číke al. 2019, 47–48). Following he comple ion o he
a chaeological and epig aphic su eys o he su ounding a ea in sp ing 2019, he F on ie
S udies will ocus on he de ailed in es iga ion o selec ed geomo phological, clima ic, and
en i onmen al phenomena and cul u al -his o ical ime ho izons. The pe iod in which he
Oh id egion expe ienced an unp eceden ed lou ishing hanks o i s in eg a ion in o he
s uc u es o he Roman wo ld, symbolized by he Via Egna ia unning h ough he a ea no h
o he lake, plays asigni ican ole in his con ex (Ke schbaum– Ve čík 2023, 71). Apa
om his, howe e , he e is e y li le a chaeological e idence. Se e al Impe ial and La e
STUDIA HERCYNIA XXVIII/2, 191–204
192 STUDIA HERCYNIA XXVIII/2
An ique nec opolises a e known om exca a ions, such as he si es a ound he monas e y o
S . Naum and in he illages o Go no and Dolno Lakoče ej, Reči sa and Volino (Ve číke al.
2019, ab.2). Howe e , ha dly any sys ema ic a chaeological in es iga ions ha e occu ed
ou side he owns o Oh id and S uga. The ew excep ions a e he wo nec opolises in O -
o nik (Ris eski 1989) and Delagožda (Bi ako a -G ozdano a 1994), loca ed on a e ace
abo e he S uga Plain, nea which illa us icae we e unea hed. The e o e, in 2021 and 2022,
an a chaeological, a chaeobo anical and biomolecula in es iga ion was conduc ed on asi e
iden i ied by p e ious a chaeological and geophysical su eys nea he p esen -day illage
o Leskoec. In his epo , we wan o p esen he geophysical in es iga ion, a chaeological
exca a ions, and he i s esul s o he ma e ial s udies.
Fig. 1: Topog aphical map in ascale 1:25 000 wi h ma ked S . Pe kina Ni a (OH_ID 2085). The si e is
placed in an ele a ed posi ion o e looking he Oh id Valley. Map by P.Tušlo á.
The in es iga ed si e lies in No h Macedonia, abou 4.5 km no h -eas o Oh id, on asou h-
-eas e n edge o Leskoec illage. I is placed on an ele a ed e ace abo e he Oh id allu ial
plain on a oo hill o he no he n p omon o y o he Galičica moun ain ange. In i s di ec
icini y we may ind chapel o S . Pe ka and achu ch o S . Thomas he Apos le (Fig. 2).
The si e is lis ed in agaze ee o a chaeological si es o No h Macedonia as S . Pe kina
Ni a, unde he en y o he Leskoec illage (AKM 1996, 249). In AKM i is desc ibed as Roman
pe iod nec opolis whe e se e al g a es we e ound du ing ag icul u al ac i i ies. One o he
g a es was supposed o con ain wo ce amic essels and ab onze coin. The loca ion o he si e
p o ided in AKM is, howe e , misleading, placing i 3 km no h -eas o he illage, in adis-
ance which ei he eaches o e o ano he illage (Kosel) o in o hea ily o es ed a ea in he
moun ains, whe e ag icul u al ac i i ies could no ha e aken place. The only possibili y o
193P. TUŠLOVÁ– M. VERČÍK– P. ARDJANLIEV– M. VÁGNER– J. HAVELKA– D. ŠÁLKA
i s associa ion is wi h he chapel o S . Pe ka si ua ed abou 1.2 km sou h -eas om he cen e
o he illage, and nea by land, known locally as S . Pe kina Ni a (Fig. 1).
The si e o S . Pe kina Ni a is loca ed in cul i a ed ields, which a e placed in aslope sou h
o he men ioned chapel and he chu ch. The ield is di ided in wo by adi oad, o abigge –
eas e n– and asmalle – wes e n– pa . Du ing su ace su ey in 2019, which was ca ied ou
wi hin he ame o he p ojec F on ie S udies, se e al she ds we e ound on he eas e n pa
o he ield. Despi e being mos ly hea ily agmen ed and undiagnos ic, he po e y iden i ied
p obable loca ion o he si e. The sca e was gi en iden i ica ion no. OH_ID 2085 acco ding
o he p ojec s numbe ing sys em. The p elimina y ch onology based on he su ey was es-
ablished in o he ange o he La e Hellenis ic / La e Republican o Roman Impe ial pe iod
(Ve číke al. 2019, 34, ab.2).
In 2021 ageomagne ic p ospec ion was conduc ed on he eas e n hal o he ield (Pl. 4/1–5).
I iden i ied se e al ea u es concen a ed on i s no he n pa , including s one s uc u es.
The ollowing yea , asmall ench o 4 × 4 m was placed on he iden i ied s uc u e o assess
i s cha ac e is ics and ch onology (Fig. 2).
Fig. 2: Field o S . Pe kina Ni a wi h ma ked 4 × 4 m ench exca a ed in 2022. The building a he
no h is he chu ch o S . Thomas he Apos le, he chapel o S . Pe kina Ni a placed along he
oad is hidden unde he ees. Map by P.Tušlo á.
194 STUDIA HERCYNIA XXVIII/2
GEOPHYSICAL PROSPECTION
In 2021, a eam om he Depa men o A chaeology and Museology a Masa yk Uni e si y in
B no, Czech Republic, ca ied ou ageophysical su ey a he si e.
1
The p ospec ion ocused on
he eas e n pa o he ield. Combina ion o magne ome y (MAG)2 and g ound -pene a ing
ada (GPR)3 was used. The ield was p ospec ed in Augus , when he ege a ion was low and
he ield walkable. The GPR su ey was conduc ed on 4526 sq m (Pl. 4/1, 2), while MAG on
6243 sq m (Pl. 4/2, 3); he wo measu emen s o e lapped (Pl. 4/5).
Se e al linea ea u es we e iden i ied in he no h -eas , a he mos ele a ed pa o he
su eyed ield, close o he mode n oad. They can be in e p e ed as possible emains o walls
su ounded by s one deb is and se e al pi -like ea u es. The GPR su ey u he showed ha
as ongly e lec i e inhomogeneous laye s e ches owa ds he sou heas . I is possible ha ,
in addi ion o ubble, i is abed ock ising close o he su ace (Pl. 4/5). In he cen al and
sou he n pa s o he in es iga ed a ea, he magne og am shows ex ensi e and s ongly mag-
ne ic su ace anomalies ha can be in e p e ed as p obable pedological/geological s uc u es
(e.g. ou c ops o magne ic ocks, see Pl. 4/3). Thei p esence p e en ed he iden i ica ion o
any o he ele an anomalies ha could be in e p e ed as po en ial a chaeological s uc u es.
Bo h geophysical me hods, MAG and GPR, con i med he linea ea u es on he no h o be
emains o walls, placed abou 30–40 cm below he e ain. Ne e heless, hey also p o ided
di e en g ound plans, only wi h se e al walls o e lapping. The si ua ion in he es o he
ield could no be e alua ed due o he pedologically/geologically un a ou able condi ions.
EXCAVATION
Based on he geomagne ic esul s a es ench was placed on ac ossing o se e al walls,
iden i ied by bo h measu emen s, bu especially isible on he GPR which ha e shown mo e
cohe en s uc u e o along house wi h a oom in i s sou h -eas e n co ne . The ench meas-
u ed 4 × 4 m and i was placed o co e inne and ou e space o he p esumed house (Fig. 3).
1 The da a p esen ed he e we e ob ained and p ocessed by Michal Vágne , Da id Šálka, and Jan
Ha elka.
2 The magne ic su ey o he si e was ca ied ou wi h a luxga e magne ome e LEA MAX (Eas e n
A las) wi h 10 Fe ex CON650 (FOERSTER) p obes. The densi y o he magne ic measu emen s was
0.5 m on he X -axis and 0.1 m on he Y -axis (measu emen di ec ion). Measu ed da a we e collec ed
wi h spa ial in o ma ion om he GNSS ecei e and p ocessed in LEAD2 so wa e. Magne og ams
(in uni s o nano esla nT) we e smoo hed by a e aging. To in e p e and p esen he esul s A cGIS
P o (ESRI) so wa e was used.
3 Fo GPR su ey amul i -channel GPR Impulse Rada Rap o -45 was used, equipped wi h eigh
shielded an ennas wi h acen al equency o 450 MHz. The spacing be ween he measu emen
lines has been se a 0.085 m. The densi y o measu emen poin s on each p o ile was se o 0.1 m.
The ob ained da a we e p ocessed in he GPR Slice so wa e (Geophysical A chaeome y Labo
-
a o y/Sc eening Eagle Technologies, USA), whe e ecommended il e s (Time -Ze o Remo al,
Backg ound Remo al, Bandpass il e , Hilbe T ans o m) we e applied o ei he highligh o
supp ess he de ec ed anomalies (s uc u es) in di e en pa s o he ime slice. The esul ing
ho izon al ime/dep h slices we e expo ed in he o m o a as e o A cGIS P o (ESRI), whe e
hey we e geo e e enced o he coo dina es o he co ne s o he ele an polygon. The iden i ied
anomalies– po en ial a chaeological s uc u es– we e subsequen ly in e p e ed, ca ego ised, and
digi ised in o a ec o plan.
195P. TUŠLOVÁ– M. VERČÍK– P. ARDJANLIEV– M. VÁGNER– J. HAVELKA– D. ŠÁLKA
Fig. 3: Resul s o he GPR su ey a he 40 cm dep h wi h isible s one ounda ions and ma ked
4 × 4 m ench exca a ed in 2022. Map by P.Tušlo á.
The en -day exca a ion was conduc ed in Augus 2022 on ha es ed ield. The ench was placed
in agen le slope, a he heigh o 785 m.a.s.l., sligh ly below he highes peak o his speci ic
hill. The 4 × 4 m ench was placed along he p esumed wall, esul ing in NW–SE o ien a ion
o he ench (Fig. 3), wi h ele a ion di e ence o 45 cm on he ou me es, sloping om
SW o NE. All inds, compound o oo iles, po e y, glass agmen s and small me al objec s,
we e collec ed and documen ed, so we e he soil samples, cha coals, animal bones and snails.
The opsoil [SU01] was abou 10–20 cm hick, depending on he slope inclina ion, and i
yielded agmen ed po e y mixing Roman and ecen she ds, se e al pieces o i on including
anail, ahook, and ashaped b onze pla e (Tab.1: 2–4, 6, Fig. 8: 1–2, 4). Bellow he opsoil ap-
pea ed yellowish laye mixed up wi h g a el [SU02], i was o simila hickens as he opsoil
(10–20 cm). Few po e y agmen s, an undiagnos ic agmen o aglass essel, and an i on
nail we e ound he e (Tab.1: 1, 5, Fig. 8: 3). Again, we can assume ha Roman and mo e
ecen ma e ial was mixed in his laye .
The [SU02] was si ing on as one laye . The s ones we e bigge and mo e massi e on he
SW pa o he ench, while he NE pa ea u ed smalle -sized s ones. The isual di e ence
in he s one size and dis ibu ion was enhanced by a egula line o s ones unning ac oss
he ench, which u ned ou o be solid wall pa allel o he edge o he ench, di iding i
in o wo pa s (Fig. 4).
Subsequen exca a ion ully unco e ed he wall and disclosed– on he wes along he
no h -wes e n edge o he ench– ano he wall pe pendicula o he i s one and joining i .
Bo h walls we e made o d y laid ubble s ones mixed up wi h soil con aining some ce amic
agmen s. Bigge s ones we e used o he wall’sedges, smalle ones we e placed in i s inne

196 STUDIA HERCYNIA XXVIII/2
pa . Bo h walls we e 55–60 cm wide and p ese ed o he heigh o abou 60 cm, placed di ec -
ly on he bed ock, which was sligh ly hewn o accommoda e he ounda ion: abou 5–12 cm
in dep h, and up o 10 cm in wid h a ound he walls (Fig. 5). The wes - unning wall ended
sho y be o e he edge o he ench, possibly e ealing he en ance in o he s uc u e– an
84 cm wide opening was cap u ed wi hin he ench. Signs o bu ning we e ound in his pa
o possible en ance, ea u ing high concen a ion o cha coal.
The wall unning in he NE -SW di ec ion di ided he ench in wo pa s. The wes e n
one was illed by big -size s ones [SU05] si ing on asoil laye [SU07], co e ing on i s u n he
bed ock, and p obably sugges ing i o be he indoo pa o he house co e ed by des uc ion
deb is. The eas e n pa , wi h small size s ones [SU03], placed on laye wi h e y li le inds
[SU04], seems o be he ou doo pa o he house, p obably some ou e pa emen . They we e
bo h co e ed by he same wo uppe laye s– b own opsoil [SU01] and he yellow g a el laye
[SU02] paced di ec ly o e he s ones (bo h [SU003] and [SU005]).
Fig. 4: The exca a ed ench in di e en le els o unco e ing. On he le , a e he emo al o
[SU01] and [SU02]. On he igh – SW pa : si ua ion a he bed ock wi h one dolium base in
si u and se e al sca e ed agmen s on he sou h o i ; NE pa : deepe cleaning on he s ones
[SU03]. Digi alized by D. Mildo á.
197P. TUŠLOVÁ– M. VERČÍK– P. ARDJANLIEV– M. VÁGNER– J. HAVELKA– D. ŠÁLKA
Fig. 5: On he le , inal s age o he exca a ed ench, SW pa : a e he po e y emo al and
cleaning o he p obable en ance a he NW co ne , NE pa : hal o he pa exca a ed o he
bed ock wi h an unco e ed pi , cu in hal . On he igh , sec ion o he pi a e emo al o he
wo uppe laye s [SU01] and [SU02]. D awn by N. A anasoska V hel, digi alized by D. Mildo á.
THE WESTERN PART OF THE TRENCH
Amassi e s one laye abou 60 cm hick [SU05] was ound unde he yellow laye [SU02]. The
s ones we e mixed up wi h soil and agmen ed ma e ial including wo glass agmen s da ed
in o he Roman pe iod and one i on nail inse ed in o an i on ing (Fig. 8: 5, 7–8). Unde he
massi e s one laye , a loo le el abou 10 cm hick, was iden i ied [SU07], placed di ec ly on
he bed ock. The laye [SU07] con ained la ge amoun o po e y wi h ela i ely low deg ee
od agmen a ion, including pa s o se e al dolia; he e we e no small inds. Fou di e en -
ly shaped dolia ims we e dis inguished (Fig. 9: 1–4). The ab ic o wo o hem was clea ly
ecognizable, enabling hei iden i ica ion wi hin di e en a eas o he ench. F agmen s
o he dolium Fig. 9: 3 we e sp ead o e he cen al pa o he ench jus abo e he bed ock,
whe e i c ea ed small pile composed o body agmen s only; se e al o he agmen s o he
same dolium, including pa s o a im, we e ound nea by in apos hole/dep ession o ixing
he dolium (?) (Fig. 7– all o he isible po e y she ds).
198 STUDIA HERCYNIA XXVIII/2
The base o ano he dolium, Fig. 9: 5, was ound in si u, jus no h o he p e ious one, wi h
lowe pa o he base s ill sunken in o he bed ock (Figs. 4 bellow and 7). Se e al body
agmen s o he same essel we e ound collapsed inside he base oge he wi h some big-
ge pieces o oo iles. The dolium p ese ed hick da k laye o o ganic ma e inside he
base, which was hea ily bu ned, making he essel e y agile when eco e ing i om he
soil. F om abou 250 po e y agmen s ound in [SU07], se e al o he s we e also coa ed wi h
ablackish laye on hei inside, ei he con inuously o spa sely. These we e mos ly, hough
no exclusi ely, coa se wa e she ds.
The bed ock was eached a he dep h o 90 cm in he deepes SW co ne o he ench. I
consis ed o whi ish (lime-)s one, excep o he sou he n pa o he ench, whe e he s one
was so e , o eddish in (Fig. 7). In his eddish pa o he bed ock apos hole was ound,
placed abou 36 cm wes om he wall. I was 20 × 22 cm in diame e and i s dep h a ied om
5 o 12 cm. Ano he wo ea u es iden i ied as pos holes we e loca ed in he whi ish pa o
he bed ock. One was al eady men ioned abo e, con aining emains o he dolium Fig. 9: 3
Fig. 6: Eas e n, sou he n, and wes e n sec ions o he exca a ed ench ( om op o bo om). D a-
wn by N. A anasoska V hel, digi alized by D. Mildo á.
199P. TUŠLOVÁ– M. VERČÍK– P. ARDJANLIEV– M. VÁGNER– J. HAVELKA– D. ŠÁLKA
(d. 32 × 40 cm, dep h 17 cm). Ano he one was possibly placed in be ween he wo (d. 20 × 27 cm,
dep h 13 cm). These h ee ea u es– ega dless o hei sligh ly di e en dimensions– a e all
aligned and pa allel o he s one wall (Fig. 5).
THE EASTERN PART OF THE TRENCH
The eas e n pa o he ench was smalle , abou 120 cm om he edge o he ench o he wall.
Smalle -sized s ones accumula ed he e, wi h occasional la ge s ones, mixed up wi h soil and
a iable ma e ial [SU03]. A e exca a ing h ough he wo uppe laye s [SU01] and [SU02],
and cleaning on he op o he s one laye [SU03], we con inued in asmalle a ea, cu ing he
4 m ench in hal (Fig. 5). The bed ock was again eached a 90 cm in he deepes SE co ne
o he ench, cu ing h ough alaye [SU04], which was no de ec ed in he wes e n pa . I
is b ownish in colou , con aining cha coal, agmen s o a chi ec u al ce amics, pebbles, and
one i on hook (Fig. 8:6). The ca. 20 cm hick laye was si ing di ec ly on he bed ock.
Exac ly in he middle o he ench leg h we unco e ed api , abou 90 cm wide, and 40 cm
deep, cu di ec ly in he bed ock. The ill o he pi was simila in composi ion o he laye
[SU04], bu he soil con ained e en highe amoun o cha coal and i was isibly sloping o-
Fig. 7: Pho o o he si ua ion d awn on Fig. 4 le . The dolium sunken o he bed ock is ba ely isible.
The eddish colou ing o he bed ock can be seen nea he sou he n wall. Pho o P.A djanlie .