ANcpr MoRrLro, NoRBERT HeNBr & EspERANZA MenriN(eds.)

LNNITES XX

XX CONGRESO NNI|ER.NACNONAt

Dtr ESTUDNOS SOtsRE LA F'RONTER.A RON,IIANA

X' lNTERA\TAzONAL CONGRESS OF" RoA4AN F.RoNzER STUDTES

Le6n (Espafia), Septiembre, 2006

Congreso celebrado bajo la presidencia de honor de su Majestad el Rey de Espafiay con el apoyo de la Universidad de Le6n

At.jos de GladiusT3

(Vol. 3)

CoNseJo SupnruoR DE INvnstlcRcroNEs CrpNrirrcasIxsrrruto Hrsr6ruco HoppMpysR. INsrrruro DE ArqunoloclR op MBnron

EorcroNr,s PoupEMoMadrid, 2009

ROMAN HORSE HARNESS FITTINGSFROM BURGEIUAE

IveN RnovnN-LIveTe

Numerous Roman finds have been discovered on the site of Novi Banovci since the 19'h

cenrury. The village is situated on the Danube river, in the southwestern part of the Serbian

province of Vojvodina. The large number of finds is easily explainable considering the fact that

the village occupies the spot of a Roman frontier fort.

Before the collapse of the Austro-Hungarian Empire the archaeological finds from that area

were usually dispatched to the Archaeological department of the Croatian National Museum.

Nowadays, due to that fact, agreat number of archaeological finds discovered on the sites of the

Srijem region, as this part of Vojvodina is called, are kept in the Archaeological Museum in

Zagreb. Among them is a fairly large collection of Roman finds from Novi Banovci, i.e. Burgenae.

Some of them have been published since, but the majority still awaits a thorough analysis and

publication. Considering the extent of that collection, and the limited amount of space for this

paper, obviously a selection had to be made. During the ROMEC congress held in Budapest in

2005,I presented the belt fitting finds from Burgenae and for this occasion, I have chosen to

presenr the finds of horse harness fittings, which can be attributed with a high level of certainry

to the Roman military horse harness.

There is no need to present in too much detail the history of the site and the military units

that garrisoned Burgenae since this topic has been extensively discussed in several publications

(Ljubi6, 1879:99-100; 1889: 62-63; BrunSmid, 1895: 172-180; Frohlich, 1892:40-44; Patsch,

1899: 1062; Brunbmid, 1901: 156-168; Radn6ti & Bark6czi, l95I: 215-216; Dimitrt jevi|

196l: 93-94; Klemenc, 196I: 22; Pileti6 & RaIi e, 196I: 87; Klemenc, 1 963: 66; Pileti6, 1972:

7-14; 1986: 138-140; 1989: 82-85; Lor incz,2001 : 6I , 63, 66, 59,72, 75,80, 82), and I wi l l

therefore just offer a short overview. \Mhile there is no much doubt that the fort was built during

the Flavian period at the latest, nothing is known for sure about its garrison in the first decades

of its existence. Some authors presume that the cohors II Asturum et Callaecorum might have been

the first unit stationed there, but this assumption is far from being certain (Spaul, 2000: 81;

Lorincz,2O0I: 29, 63). The same could be said about the coltors V Gallorum, which might have

been in Burgenae duringTlaiant reign (Spaul, 2000: 168-169; Lorincz, 200I: 34-35,80, 82).

The ala I ciuium Romanorumisthought to have garrisoned Burgenae between 118 or 119 and

138 AD, an assumption which seems quite likely (Spaul, 1994:85-86; Lorincz,200I: 18, 88-

8e).

1 500 LIMES )C( Gladius, Aneios 13, 2009

Although there are many uncertainties about the garrisoning units during the early history ofBurgenae, i.e. the Flavian and Tfaianic periods, most aurhors agree that the cohors I Thracumciuium Romanorum pia fdelis was stationed in Burgenae after 138 AD. It seems that it remainedthere for a very long time, probably until the second half of the 3'd cenrury AD, when it wasmoved to Bassianae (Radn6ti & Bark6czi, I95l 215-216; Dimitriievi6, 1961: 94; Klemenc,196l : 22; Du5ani i , 1968: 96-97; Eadie, 1977:218; Visy, 1988: 130; Spaul , 2000: 36I-352;Ltirincz, 2001: 42,90,99-101). The older authors believed that it could have been replaced bythe coltors III Alpinorum after the reign of Marcus Aurelius, bur that seems quite unlikely(\7agner, 1938:85-86; 189-190; Dimitri jevi6, 195511957:299-300; Klemenc, 1953:66). Thenames of some units are recorded for the 4'h century AD as well: one detachment of the legio VIouia, the equites Dalmatae and the cuneus equitum Constantianorum 1 (Dimitrijevi6, 196l:94;Klemenc, 196l:22;Yisy, 1988: 130). As one can see, numerous units of the Roman army arerecorded or believed to have been stationed in the Burgenae fort during almost four centuries ofits existence.

The small copper-alloy phalera (cat. 1) according to Bishops rypology of Roman cavalryequipment, is akin to his functional typ. 2f while the front side of this phalera coffesponds toBishopt decorative type 2a. This type of phalerawas most probably associated with a pendant andserved no functional purposes except as a pendant mount (Bishop, 1988: 94-95, 135-141). It canbe roughly dated to the 1" century AD, probably starting from the Claudian time to the early 2"dcentury at the latest (Bishop & Coulston 1993:105; Rabeisen, 1990: 80, 84-85; Deschler-Erb,1999: 59; Chapman, 2005: 132-133).

There are four, more or less well preserved junction loops among the harness fittings fromBurgenae. The fragment (cat.2) probablybelongs to Bishopt rype 1, i.e. the spectacle rype, butthe subtype cannot be determined. The rypes of the other three junction loops are far easier todetermine. There is a nicely preserved junction loop of 2atype (cat.3), the largest specimen ofrectangular shape with a broad decorated band at the upper end (cat. 4) is related to the type 5e,while the simple rectangular junction loop (cat. 5) could be determined as rype 8h. (Bishop,1988: 100). They can all be dated to the 1" and early 2"d century AD or even somewhat later(Deschler-Erb, 1999: 61; Chapman, 2005: 133-136).

The copper alloy strap terminal (cat. 6) is also a very common item. This specimencorresPonds to the Bishopt la rype and can be dated to the 1" cenruryAD (Bishop, 1988: 103,109, 1 54-165; Deschler-Erb, 1 999: 62; Chapman , 2005: 136-137).

The strap mounts are fairly well represented in this small sample of harness fittings. The rwolarger ones seem to correspond to Bishopt 6d rype. Actually, the better-preserved one (cat. 8)definitely does, while the fragmented one (cat. 7) is certainly related. The latter one still bearstraces of niello.

The three smaller ones are related to Bishopt 6b rype. The best preserved one (cat. 9) showsclear traces of coating, most probably silvering and traces of niello, which might also be presenron the other two (cat. 10 and 11). All of them can be dated to the 1" cenrury AD (Ulb ert, 1959:94,T. 19, 1_3; Bishop, 19gg: 170_173; Deschler_Erb, 1999: G3).

The teardrop-shaped pendant (cat. 12) is a quite common item among Roman horse harnessfinds. This particular pendant with simple openwork decoration corresponds to Bishopt rype 5d.That rype of pendants can be dated from the reign of Claudius ro the Tlaianic period, perhapseven longer (Bishop, 1988: 96,98; Deschler-Erb, 1999:57).

r Notitia Dignitatum, (Pann. II) Occ. 32, 5=24, 18=37, 46.

Gladius, Anejos 13, 2009 MOESIA & DACIA

The lunate pendants are also extremely common and present a very wide range of forms. They

have been used for an extensive period of time, from the Augustan period to the early 3'd century

and the most common types are therefore hard to date precisely when one lacks an archaeological

context (Oldenstein,1976 162-164; Bishop, 1988: 98; James,2004:88). It seems that the earlier

forms were most often flat. Therefore, those two pieces would rather belong to the 2"d century

AD. The larger fragmented lunate pendant (cat. 13) is somewhat remindful of Bishopt 9d rype,

but its knobbed terminal would also be an argument in favour of dating it to the 2"d century AD.

The smaller lunate pendant with knobbed terminals closing a circle (cat. 14) corresponds to

Bishops rype 9m which also seem to be more common in the later period, i.e. from the mid 2"d

century AD onward.

The thin fragmented harness fitting (cat. 15) belongs to a type which is variously described as

heart-shaped or pelta-shaped and which is quite common. Its shape is related to the peltiform

fittings common from the mid 2"d century AD onwards but it would seem that such larger flat

specimens are earlier in date and that they can be dated from the second half of the 1" century

AD to the mid 2"d century AD. Although many found specimens bear traces of coating, either

tinning or silvering and niello, there are no such traces on this fitting (Unz, I97I: 50, Kat. 54;

Oldenstein, 1976:181-182,Kat. 541-642; Rabeisen, 1990: 84, 88; Deschler-Erb, 1999 58-70;

Radman-Livaja, 2004: I 08-1 09).

The rwo openwork copper alloy leaf-shaped pendants decorated with volutes (sometimes such

pendants are also referred to as heart-shaped pendants) find analogies among late 2"d and 3'd

century material (Oldenstein, 1976:130, Kat. 229-233;James, 2004:90, Cat. 227). The larger

specimen (cat. 16) is fragmented and is missing its fastening loop as well as its end knob, but the

smaller piece (cat. 17) is well preserved and has a fastening loop, which was probably meant for

a hinge. Its end knob is shell-shaped.

The small peltate stud with volutes and an upper terminal in the form of a fleur-de-lys (cat.

18) is one of the many variants of this very common type of fitting. Such studs have been used

since the mid 2'd century AD, and their use continued for a long time, till the 4'h century AD

(Oldenstein, 197 5: I78-I81; Aurrecoechea, 1996: 99, I03-105).

The rwo small fittings (cat. 19 and 20), described as leaf or almond-shaped fittings with a

knob end, are not uncommon among 3'd century harness fitting finds, as witnessed by finds from

Zugmantel, Sisak or Tbutoburgium (Schleiermacher,2000: 1S8. T. 8,22; Radman-Livaja,2004:

I17 , Kat. 530; Radman-Livaj a, 2005: 941, Cat. 42-44).

Three copper alloy pieces from Burgenae are most probably bridle mounts.

The shape of the lentoid bridle mount (cat.21) is remindful of some bridle mounts found in

Pannonia, namely from the wagon graves of Kozdrmisleny and Baldca in Hungary, dated to the

Iate 2"d and 3'd century AD (Paligyi, 1997: 467, Fig. 6.70, 6.73). The oval bridle mount with

four openings (cat. 22) is similar to one specimen from Zugmantel, which could also be dated to

the 3'd century AD (Schleiermacher, 2000: 170, I80, T. 5.8). The small elaborately shaped bridle

mount (cat.23) is similar to one bridle mount from Dura Europos (James, 2004:99, Cat. 349),

and its shape corresponds closely to one bridle mount found in Sisak (Radman-Livap,2004: I78,

Kat. 540).The fragmented openwork phalera (cat. 24) is not unlike some specimens from Carnuntum

(Junkelmann, 1996:84). The peltate decoration is quite rypical for horse harness phalerae dated

to the period between the late 2"d and mid 3'd centuryAD (Oldenstein, 1975:236;Btshop &

Coulston, 1993: 157; James, 2004: 96). One should not exclude the possibiliry that it might

originally have served as a snaffle bit cheek-piece.

1 5 0 1

r502

I

Gladius, Anejos 13, 2009LIMES )O(

wW

Hffid;h wl''..,1i1 Bl::,,:ii:jl W1 , , r ; l W; , . :1I W

l#wj

Y/

ffia.$ffi

11

I

wflTIIW

W

YfItun

t$c

d

n

aH

W

Jffilri[ilfi

t

10

FIc. 1. Catalogue (drawings by Miljenka Galic)

Gladius, Aneios 13, 2009 MOESIA & DACIA

The peltate phalera with its three junction loops still in place (cat. 25) must have served as abreast junction, with one strap running downwards, joining the saddle strap and the remainingtwo joining the shoulder straps. Although I was unable to find direct analogies for this rype ofphalera, its outlook clearly points out to a Roman design. The combination of pelta motives andjunction loops, which, as it seems, went out of use by the beginning of the 3'd century, wouldrather indicate a 2"d century or perhaps an early 3'd century dating.

The shell-shaped fitting (cat. 26) was believed to be quite common in the western part of theempire, i.e. in Gaul and the Germanies but recent finds show that this type was actually quitewidespread and is encountered from Britain till Pannonia, as well as in Italy and MauretaniaTingitana (Oldenstein, 1976: I87-I88, Kat. 696-703; Gschwind , 1998: 116-120, 130-1 32). Asfar as dating is concerned, they were certainly in use in the second half of the 3'd century AD, andit is quite plausible that they were still used in the first decades of the 4s century AD.

The two fragmented spurs (cat. 27 and 28) belong to a later period. Despite their state ofpreservation, they can be fairly easily identified as rivet spurs of the Leuna C ryp., which is oftenencountered on Pannonian sites. According to dated finds, that rype was used during the 4'hcentury AD (Giesler, 1978: lI-I2,22-32).

Obviously, one cannot expect far-reaching conclusions since those are only stray finds lackingany archaeological context, but their study can nevertheless give us an interesting insight aboutthe presence of the Roman military units on that site. This is undeniably a very small sample butit is interesting to point out that, from a chronological point of view, those various finds ofRoman horse harness fittings cover a span of over 300 years. One can see items from the 1"century, like the strap mounts or some junction loops as well as some finds dated to the late 3'dor the 4'h century AD, like the spurs or some studs, and in between finds quite common for the2"d and most of the 3'o century AD.

Obviously, those finds do not imply that cavalry units were stationed in Burgenae from the 1"to the 4'h century, but they nevertheless show that horses were present in the fort on a regularbasis. As far as we know, one ala, the ala I ciuium Romanorum was probably stationed in Burgenaefor several decades in the first half of the 2"d century, while the Notitia Dignitatura records rwocavalry units stationed there in he 4'h century AD: the equites Dalmatae and the cuneus equitumConstantianorum. At least some of the finds might be related to those cavaby unirs, but it mustbe noticed that even infantry units had pack animals and their officers also had horses, whichwere probably similarly equipped as the mounts of cavalry troopers. As far as the first century isconcerned, the lack of reliable data about the garrisoning units does not exclude the possibilirythat a cavaby unit might have spent some time in Burgenae as well.

As always, stray finds give us as much insight as they contribute to raise new questions. Onecan only hope that future finds, especially epigraphic ones, and perhaps new excavations, will giveus more reliable data about Bargenae and its garrison.

Cernrocun

l. Phalera, length - 38 mm, width -36 mm (Fig. 1, 1).

2. Junction loop, length - 26 mm, width - 12 mm (Fig. 1,2).

3. Junction loop, length -7I mm, width - 12 mm (Fig. 1,3).

4. Junction loop, length - 54 mm, width - 20 mm (Fig. l, 4).

r503

r504 LIMES )O( Gladius, Anejos 13, 2009

WAWffi'*

A/".."r?j N'', \\'* p,;,,f

\ ' . , /\ , ' " i\ /sL2

m\' ,. 'il\y

15

i

L4

%wwHH

&

IIilWII

16

\;7W VW \cW2 0L91.8

ri''f\)\

ffi2 L

22

FIc. 2. Catalogue (drawings by Miljenka Galic)

.--''l ffi tr'$luwl k m1 2 3

Gladius, Anejos 13, 2009

2 7

m7//A

Wu//a

WWMw{-nw

Ftc. 3. Catalogue (drawings by Miljenka Galic)

r506 LIMES )C( Gladius, Aneios 13, 2009

5. Junction loop, Iength - 43 mm, width - 8 mm (Fig. 1' 5).

6. Strap terminal, length - 51 mm, width - 12 mm (Fig. 1,6).

7. Strap mount, length - 81mm, width - 16 mm (Fig. 1,7).

8. Strap mount, length - 80 mm, width - 16 mm (Fig. 1, 8).

9. Strap mount, length -36 mm, width - 11 mm (Fig. 1' 9).

10. Strap mount, length -29 mm, width - 9 mm (Fig. 1' 10).

11. Strap mount, length -36 mm, width - 10 mm (Fig. 1, 11).

12. Pendant, length - 60 mm, width - 37 mm (Fig. 2, 12).

13. Lunate pendant, length - 52 mm, width - 51 mm (Fig.2, l3).

14. Lunate pendant, length - 36 mm, width - 26 mm (Fig. 2, 14).

15. Flat peltiform fitting, length - 35 mm, width - 40 mm (Fig. 2, l5).

16. Openwork pendant, length - 47 IDITI, width - 47 mm (Fig. 2, 16)

17. Openwork pendant, length - 46 mm' width - 38 mm (Fig.2, 17).

18. Peltiform stud, length -24 mm, width -27 mm (Fig.2, 18).

19. Almond-shaped stud, length - 31 mm, width - 12 mm (Fig. 2, I9).

20. Almond-shaped stud, length - 29 mm, width - 15 mm (Fig.2,20).

21. Bridle mount, length - 31 mm, width - 15 mm (Fig.2,21).

22. Brrd\e mount, Iength - 33 mm, width - 33 mm (Fig, 2, 22).

23.Brid\e mount, length - 25 mm, width - 16 mm (Fig. 2' 23).

24. Fragmented phalera,length - 35 mm, width - 54 mm (Fig. 3' 24).

25. Pekrform phalera with junction loops, length - Il2 mm, width - 175 mm; phalera:

L- 49 mm, W - 45 mm; length of junction loops, respectively: 68 mm,59 mm,7l mm

(Fig.3,25) .

26. Shell-shaped stud, length - 27 mm, width - 29 mm (Fig. 3,25).

27. Spur,length - 39 mm, width - 69 mm (Fig. 3' 27).

28. Spur, length -37 mm, width- 51 mm (Fig. 3' 28).

BrsLrocRApHv

AunnecoECHEA FpnNANoEZ, J. Q996): "Bronze studs from Roman Spain", Journal Rom.

Military Equip. Stud. 7,97-146.

Brsuop, M. C. (1988): Caualry equipment of the Roman arm)/ in thefirst century,'4D, BARInt.

Series 394, Oxford.

BtSHop, M. 6. CoulsroN, J. C. N. (1993): Roman Military Equipmenr, London.

BnuNivro, J. (1895): 'Arheololke biljelke iz Dalmacije i Panonije", Viestnih Hruatskoga

Arheololkoga Druitua, Noua Snija I, 148-183.

BnuNiurD, J. (1901): 'Arheolobke biljebke iz Dalmacije i Panonije IY", Viestnik Hruatskoga Arh.

Druitua, Noua Seriia V, 87 -168.

Gladius, Anejos 13, 2009

CHepuRN, E. M. (2005): A Catalogue of Roman Military Equipment in the National Museum of

Wales, BAR British Series 388, Oxford.

DpscHr-pR-ERB, E. (1999): Ad arma! Riimisches Milittir des 1. Jabrhunderts n. Chr. in Augusta

Raurica, Forschungen in Augst, Band 28.

Drurrruynr,ri, D. (195611957): "Rad zemunskog Narodnog muzeja na rimskom limesu u

Sremu", Starinar, Noua serija VII-VIII, 299-308.

Drurrruyrvri, D. (1961): "Nekoliko podataka o rimskom limesu u Istodnom Sremu", Limes u

Jugoslauiji I, Zbornik radoua sa simposiuma o Limesu 1960. godine, Beograd, 93-107.

DuinNri, S. (1968): "Rimska vojska u istoinom Sremu, Beogradski Univerzitet", Zbornik

Filozof hog Fakubeta X-L, 87 -l13.

ERorn, J. \f. 0977): "The development of the Pannonian frontier south of the Drave", Ahten des

XI. Internationalen Limeskongresses (Szdkesfeh&vir, 1975), Budapest, 209-222.

FnOHrrcu, R. (1 892): "Eml6kek ds leletek. Acumincum viddke ds a r|,gi Pannoniai Limes",

Archaeo logiai Ertesitii XII, 34-44.

GrnsLnR, U. (1978): "Jiingerkaiserzeitliche Nietknopfsporen mit Dreipunkthalterung vom Typ

LEUNA" , SaalburgJahrb. 35, 5-56.

GscHwrND, M. (1998): "Pferdgeschirrbeschlage der zweiten Helfte des 3. Jahrhunderts aus

Abusina/Eining", Saalburg Jahrb. 49, II2-138.

Jevns, S. (2004): Excauations at Dura-Europos 1928-1937, Final Report VII, The Arms and

Armour and other Military Equipmenr, London.

JuNTEI-vaNN, M. (1996): Reiter wie Statuen Aus Erz, Mainz am Rhein.

KrEltrNc, J. (1961): "Limes u Donjoj Panoniji, Limes u Jugoslaviji I", Zbornik radoua sa

simposiuma o Limesu 1960. Godine, Beograd, 5-34.

KrEunNc,J. 0963): "Der Pannonische Limes in Jugoslawiei', Ar/teoloiki radoui i raspraue 3,55-68.

Lyuui, S. (f gZg) : "Rittium (Surduk)" , Viestnih Hruatskoga Arh. DrustVaIl4,97-1U.

L;unri, S. (tSSg): "Rimske starine u Novih Banovcih", Viestnih Hruatshoga Arh. DrustVaXll2,

1889, 62-63.

LORrNcz, B. (200l): Die rtimischen Hilftruppen in Pannonien wiihrend der Prinzipatszeit, Tbil I:

Die Inschrirten, \Wien.

Or-opNsrEIN, J. G976): "ZLrr Ausriistung r<imischer Auxiliareinheiten", Ber. RGK 57, 5l-284.

Pnr-Acrr, S. (1997): "Rekonstruktionsmoglichkeit des Zaumzeuges aus dem Wagengrab vonKozd.rmisleny (PannoniA)", W. GnoENMAN-vAN

'W'nnrEpuNGE, B. L. veN BeEK, W. J. H.

'Wrlrevs S( S. L. 'WyNre (eds.): Roman Frontier Studies 1995. Proc. XVI', Inturnat. Congress

Roman Frontier Stud., Oxbow. Monogr. 91, Oxford,467-47I.

PRTSCH, C. (1S99): "Burgenae', kE,IIJ, 1052.

Pu-E,.rri, D. & Reiri, B. (196I): "Pregled radovaVojnog muzejana Limesu od Novih Banovaca

do Sremskih Karlovaci', Limes uJugoslauiji I, Zbornik radoaa sd simposiuma o Limesu 1960.

Godine, Beograd, 87 -92.

MOESIA & DACIA r507

l 508 LIMES )O( Gladius, Anejos 13, 2009

PrrErIi, D. (1972): "Preliminarni rezultati istraLivanja rimskog kastruma Burgenae u NovimBanovcima 1971. Godine" , Wsnik Vojnog muzeja lg,7-26.

Prrprti, D. (1986): 'Acumincum i Burgenae- dva vai,nauporiSta na donjopanonskom limesu",

Odbrambeni sistemi u praistoriji i antici na tlu Jugoslautje,-Materijali XXII,-Referati XII KongresaarheologaJugoslauije (l.Jovi Sad, 1984), Novi Sad, 137-144.

Prrprti, D. (1989): "Rimske fortifikacije na teritoriji SR Srbije", Wsnih Vojnog muzeja 33,79-r04.

ResEAsEN, E. (1990): "La production d'dquipement de cavalerie au ler s. aprbs J.-C. ). Alesoa(Alise-Sainte-Reine, C6te-d'Or, France)",JournAl Rom. Military Eqaip. Stud. I,73-98.

ReoNol, A. & Bem6czr, L. (1951): "The Distribution of Tioops in Pannonia Inferior duringthe 2"d century A.D.", Acta Arch. Acad. Scien. Hungaricae r, 3-4, l9l-230.

RnoueN-LIveJA, I. (2004): Militaria Sisciensia, Nalazi rimshe uojne opreme iz Sisha u fundusuArheoloihog muzeja u Zagrebu (Finds of the Roman military equipment in Sisah in the"holdingsof the Archaeological Museum in Zagreb), Musei Archaeoiogici Zagrabiensis Catalogi .,Monographiae, Vol. 1, Zagreb.

ReoueN-Lve1R, I. (2005): "Finds of Roman military equipment from Teutoburgiltm", Z. ytsv(ed.): Limes XIX. Proc. XIX, Inturnat. Congress Roman Fiontier Stud. (Pdcs, 20:B), pdcs, 939-952.

SrAut, J. 0994): Ala,, Andover.

SIAUL, J. (2000): Coltors', BAR Int. Series 841, Oxford.

UtnERT, G. (1959): Die riimischen Donau-Kastelle Aislingen and Burghiife, Limesforschungen,Band 1, Berlin.

UNZ, C. (1971): "Rcimische Militdrfunde aus Baden - Aquae Helueticae", Jahresbericht I97I,Gesellschafi pro Vindonissa 1972, 4I-58.

vISv, z. (1988): Der pannonische Limes in (Jngarn, Stuttgart.

Ytsv, Z. (ed.) (2003): The Roman Army in Pannonia, An Archaeological Guide of the RipaPannonica. Pdcs.