THE POSSIBILITIES OF IDENTIFYING COINS 
FROM THE IMPERIAL MINT AT VIMINACIUM, 
ON THE BASIS OF NEW FINDS*
VLADIM IR KONDIČ
A rheološki in stitut, Beograd
One of the chief problem s is the problem of the im perial m int a t V im ina- 
cium and, to a certain ex ten t of th a t in M ediolanum . The m ajority of solutions 
proposed seem to be ra th e r superficial or even u ntenable. In spite of the 
freq u en t changes of opinion expressed by num ism atists, on these m ints and 
in spite of th e lim ited factual basis fo r th eir existence, th e problem w as offi­
cially settled, so to say, in the book by M attingly and Sutherland.
Though an incontrovertible solution of the problem is im possible w ith th e 
evidence available, a revision of it seems both feasible and necessary; th e 
follow ing facts m ay be taken as reliable starting-points.
1. I t is practically beyond doubt th a t th e A ntoniniani of P acatian w ere 
issued in V im inacium .
2. The m in t of the colony of V im inacium began to w ork in 239.
3. The historical circum stances — b arb arian attacks on th e southern p a rt 
of D anube, th e presence of num erous soldiers stationed in the fortresses along 
the lim es, freq u en t rebellions and usurpations on th e p a rt of th e legionary 
officers — dem anded prom pt financial interventions w hich, in tu rn , m ade a 
m int n e a r to the scene of action an indispensable factor.
4. A careful analysis of hoard s from th e n o rth e rn B alkans m ay detect 
the types of coins found only h ere or found here in strik in g ly g reater q u an ti­
ties th a n in other finds.
F irst, w e should note th a t th e type VICTORIA AVGG (RIC 51) of P hilip I 
has been identified by Eddy as originating from V im inacium and has been 
linked w ith the coinage of P acatian. A ccording to E ddy's convincing recon­
struction, P acatian began to issue silver in V im inacium w ith the reverse dies 
of P hilip, found allready m ade in the m int. This supposition, p referable to 
th at of M attingly th a t th e type in question should be tre a te d as a hybrid one, 
involves a beginning of the m in t's w ork earlier th an P acatian ’s accession. The 
same is suggested by a com parison of the distrib u tio n of this type in the
* A p p a ra tu s criticus, lite ra tu re an d B eogradski n a la z denara i antoninijana, 
discussion a re given fu lly in th e m ono- S e p tim ije S ev e r-V a le rija n (B eograd: M u- 
g raph p u b lish ed m ean tim e by th e a u th o r: zej g rad a B eograda, 1969) p. 185 ff.

D orchester and Smederevo hoards: the latter reaches 92 °/o of the form er. In 
the to tal of Philip’s coins from Smederevo, this type represents 4,60 °/o of the 
whole; sim ilarly is the Singidunum hoard w here the Figure is 4,58 ® /o . All 
these indications, enum erated by Eddy, as w ell as his calculation of the d istri­
bution of the m aterial discussed in B alkan and other hoards, are supported 
by our find. W ith sim ilar argum ents, Eddy also attrib u tes RIC 226, AETERNIT 
IMPER, of Philip II to Vim inacium . This type, again, is rem arkably represen­
ted in Sm ederevo and other B alkan hoards, m uch b e tte r than in hoards from 
the w estern provinces. The proportion offered by the Singidunum hoard foi 
RIC 226, AETERNIT IMPER, is still stronger: of the total of 118 Philip II coins 
in the hoard, this type num bers 17, or 14,49 °/o, w hile the figures of the Sm e­
derevo find are 319 : 22 = 6,89 °/o. Eddy’s attribution, based on the com plete 
statistical evidence, seems to us quite justified, obviously preferable to ascrib­
ing th e type to the Roman m int, w hich appears arb itrary .
The Jablanica hoard provides interesting data on the coinage of T raianus 
Decius. It contains a specimen, otherw ise unknow n, w ith the inscriptions 
(Obv.) IM P TRAIANVS DECIVS AVG (Rev.) PM TR P III COS II P P (Securitas 
seated 1., holding sceptres, w ith head leaning against 1. hand). Because of th e 
datable titles, this coin w as used by M. Vasić as a starting-point for dating 
the issue bearing the shorter obverse legend IMP TRAIANVS DECIVS AVG, to 
the end of 250 or the beginning of 251, and th a t w ith the longer legend, IM P C 
M Q TRAIANVS DECIVS AVG, to 249. Though th e series w ith the shorter 
obverse inscription is usually assigned to the R om an m int, M. Vasić allowed 
the possibility th a t it actually originates from a B alkan mint, probably the 
m oneta comitatensis, w hich could have produced this unique coin. As he is 
of the opinion th a t this period was too early for th e coining of silver-coins at 
Vim inacium , he is reluctant to accept the possibility of assigning these coins to 
the above m entioned m int. However, there exist im portant indications th a t the 
m int of Vim inacium was opened as early as in P h ilip ’s reign and — if this 
is also tru e of Pacatian, w hich is practically certain — it is quite probable th a t 
the m in t continued its production under T raianus Decius. The coin from 
Jablanica renders such a supposition very probable.
The so-called M ilan series of T raianus Decius, w ith the obverse legend 
IMPCAETRADEC/IVS/AVG, and the Consecratio series, though com paratively 
rare, are w ell represented in the Singidunum hoard. The same is also tru e 
of the Jablanica hoard (even 26 Consecratio specimens and 9 others, out of the 
total of 371) and th a t from Sm ederevo (123 coins, as com pared to Decius’ total 
of 1389). Such a distribution would indicate a neighbouring m int, the m ore 
so as the theory regarding the M ilan m int prior to the reign of G allienus 
rests upon w eak argum ents. If th e m int responsible is really to be sought in 
this area, then the only possibility — so fa r as w e can see — is Viminacium. 
This city possessed a m int for' colonial money as early as 239, which m eans 
the presence of trained men — some m em bers of the im perial m in t’s personel 
m ust have been transferred from Rome, e. g. engravers — and the necessary 
apparatus. In spite of the difference in m etal and size of the respective pro­
ductions, the existence of th e local m int in V im inacium m ust have given the 
city a considerable advantage over other possible candidates for the location 
of the im perial m int, an advantage clearly realized by Pacatian at least,

whose A ntoniniani undoubtedly originate from V im inacium . In all likelihood, 
the m inting th a t had taken place in Vim inacium u n d er th e rebel w as to con­
tinue fo r a while, at least d u rin g th e period of Decius.
T he coins of T rebonianus G allus from our hoard provide some fu rth e r 
indications for a neighbouring m int, speaking, it seems, against M attingly’s 
assum ption of the M ilan m int functioning in these years, or against his iden­
tification of the M ilan production. We shall begin w ith the specimens, o th er­
wise unknow n, w hich appear b oth in the Jablanica and Singidunum hoards:
Obv. IM P CAE C VIB TREB GALLVS
Rev. PA X AETERNA
T he type PAX AETERNA, am ong the issues of Trebonianus G allus has 
so fa r only been found on R IC 71, w ith the sam e reverse representation, b u t 
w ith a d ifferen t obverse legend, IM P CC VIB TREB GALLVS AVG (assigned 
to M ilan in RIC). According to M attingly-S utherland, th e obverse of RIC 71 
w ould be Roman, the reverse characteristic of M ediolanum . H owever, our 
specim en dem and a revision of such an attribution, and throw doubt upon the 
assum ption of a M ilan m in t u n d er Trebonianus G allus. It could hard ly be 
taken for a sim ple coincidence th at, not fa r from Vim inacium , tw o identical, 
otherw ise unknow n specim ens w ere found. A ll this speaks in favour of a sta te 
m int in Vim inacium . M. Vasić, also, surm ised a local m int in this connection, 
bu t w e th in k th a t the attrib u tio n should be m ore precise, in view of the arg u ­
m ents adduced above in favour of the V im inacian m int.
T he hoards from Jab lan ica and Singidunum list tw o m ore new specim ens 
of T rebonianus Gallus, w hich suggest sim ilar conclusions. The first is from 
Jablanica: (Obv.) IM P CC VIB TREB GALLVS, (Rev.) PA X AETERN. A ccord­
ing to M attingly-S utherland, th e obverse w ould be from Milan, the reverse 
unknow n. T hat from Singidunum reads as follows: (Obv.) IMP CAE C VIB 
TREB GALLVS, (Rev.) PIETA S AVGG (Pietas standing, w ith head 1., both 
hands raised; altar). A ccording to RIC, no A ntoninianus w ith such a com bi­
nation of the obverse and rev erse is know n. A sim ilar representation, w ith o u t 
altar, is found on the v ariety RIC 41. The reverse w ith Pietas and the altar, 
identical to ours, is described on specimens, usually derived from M ilan, w ith 
a d ifferen t obverse legend, IM P CC VIB TREB GALLVS AVG. These tw o 
specim ens, as w ell as some o th ers treated by M attingly-S utherland as com bi­
nations produced b y M ilan w ith aid of Rom an reverses actually rep resen t ra re 
coins. Taking into consideration, especially their provenance, w e th in k th a t 
th e tw o unica described m u st have originated from V im inacium . The sam e is 
suggested as probable, though not certain, for o th er types w hich occur in 
g re a te r quantities in M oesian hoards, in spite of th e ir having been assigned 
to M ediolanum . In any case, a revision of the M ilan hypothesis is required, 
and its early opening at least seem s dubious.
The last em peror w hose coins — four A ntoniniani —- are found in the 
S ingidunum hoard is V alerian. R. Gobi, in the w ell-know n p a rt V /l of the 
“A u fb au der röm ischen M ünzprägung in d er K aiserzeit”, presents a convincing 
identification and chronological arrangem ent fo r V alerian’s im perial issues 
from Vim inacium . C om pletely accepting his results, w e m ay note th a t th ey are 
corroborated by th e evidence of the Singidunum hoard. O ur No 1, w ith th e 
reverse inscription VIRTVS AVG, has been defined by Gobi as a v a ria n t of

the first V im inacian issue and dated in the m onths before G allienus was 
appointed A ugustus, i. e. before O ctober 253. The specimens w ith VIRTVS 
AVGG w ould belong to the second V im inacian issue, being coined before the 
end of th e year. It is true, th ey appear (two of them ) in our hoard w ith th e 
legend identical to those published by Gobi and others, b u t w ith a different 
representation. However, the sam e type occure on a coin from the first Vimi­
nacian issue and now here else during V alerian’s m inting. The stylistic resem ­
blance of th e portraits and th e reverses of these tw o variants, as w ell as the 
fact th a t th e reverse type of Nos 2 and 3 has been registered only on the 
V im inacian coinage, clearly locate the m int of this variety, so far unknow n. 
It m ust have been produced — in a rath er sm all quantity, judging by its 
rarity — a t the very beginning of the second issue, as it directly follows the 
reverse type of the first issue. The th ird type, w ith FIDES MILITVM has also 
been ascribed by Gobi to the second Vim inacian issue, before the end of 253. 
The last m onths of 253 would consequently represent a te r m in u s ad q u e m of 
the hoard’s burial. The in terv al betw een the outpout of Nos 1—4 and their 
arrival in th e hands of the ow ner of the Singidunum hoard may have been 
short, in view of the negligible distance betw een V im inacium and Singidunum .
O ur discussion of the problem of the im perial m in t in Vim inacium gives 
no definitive results. However, though the grounds for the m ajority of our 
attrib u tio n s m ay seem unsatisfactory or insufficient, they are supported by a 
broad set of indications, which should not be neglected. New finds will, doubt­
lessly, enrich our evidence, and provide a b etter base for fu rth er research.
M ogućnosti id en tijiko va n ja proizvoda carske ko vn ice u V im inacijum u 
na osnovu novih nalaza
P ov zetek in delni prevod tega p red av an ja je a v to r o b javil m edtem v m onografiji: 
B eogradski nalaz denaro i antoninijana, S ep tim ije S ever-V a lerija n (Beograd: M uzej 
grad a B eograda, 1969) na stran eh 47 ss, k je r sta navedeni tu d i lite ra tu ra in diskusija.