ANTHEMIUS: ROME

PRELIMINARY ANALYSIS

PART ONE: TYPES, SPECIMENS, AND DIES

Although I have catalogued many more specimens from the mint in Rome than those from the mints in Milan and Ravenna there are still insufficient numbers to justify the type of analysis employed for those other two mints. While there are two extended linked sequences involving ten and forty-four specimens, comprising four obverse and six reverse dies and thirteen obverse and ten reverse dies respectively (see part two, below), for the most part there are few linked series involving anything more than a few coins. As a result I have adopted a different format for these analyses from those analyses of the issues of Milan and Ravenna. As time progresses and I obtain more photos I hope that this will change and that a clearer picture of the order of production will arise.

I should also mention that for the Vestal hoard I have selected only a single specimen from each of the three different obverse and reverse combinations from the main die-linked sequence of cat. nos. 22-355 (i.e. 33, 145, and 354) so as not to distort the specimen counts. For Euphemia and the other mints, I have counted all available specimens.

KEY: (2s: 2o/1r*) = two specimens, two obverse dies, and one reverse die, which is shared with another RIC number (see Part Two below for links)

‘var’ indicates a type not described in RIC

‘fn’ indicates a type mentioned by RIC in a footnote

:+: = cross with pellets

/ / COMOB = 

COMOB

ANTHEMIUS

SERIES 1

Obverse: Pearl and rosette diademed, draped and cuirassed bust right (D4)

Reverse: Leo and Anthemius, facing, clad as cloaked soldiers, each holding a spear in outside hand and together supporting globus cruciger between them.

RIC NO. Obverse / Reverse

2801 (1s: 1o/1r*) DNPROCANTHE-MIVSPFAVG

SALVSREI-PV-BLICAE //  / COMOB

SERIES 2

Obverse: Helmeted bust facing, pearl-diademed, cuirassed, in right hand spear over shoulder behind head, on left arm decorated shield (KFa3)

Reverse: Leo and Anthemius, facing, clad as cloaked soldiers, each holding spear in outside hand and together supporting globus cruciger between them.

2802 (1s: 1o/1r*) DNANTHEMI-VSPFAVG

SALVSREI-PV-BLICAE //  / COMOB

SERIES 3

Obverse: Helmeted bust facing, pearl-diademed (no central jewel or trefoil), cuirassed, in right hand spear over shoulder behind head, on left arm decorated shield (KFa3)

Reverse: Leo and Anthemius, facing, clad as cloaked soldiers: Leo holds in between the figures in his right hand a shield, Anthemius holds a wreath-bearing Victory upon a globe. The emperors hold hands. Above and between is a wreath with two ties surmounted by a cross. PAS or PAX appears within the wreath.

2804var1 (1s: 1o*/1r) DNANTHEMI-V-SPFAVG

SALVSREI-PV-BLICAE // PAS // R | | M / COMOB

2804var2 (1s: 1o/1r) DNANTHEMI-VS-PFAVG

SALVSREI-PV-BLICAE // PAX // R | • | M / COMOB

2804var3 (1s: 1o/1r) DNANTHEMI-VS-P-FAVG

SALVSREI-PV-BLICAE // PAX // R | ^. | M / COMOB

2804var4 (1s: 1o/1r) DNANTHEMI-VS-PFAVG

SALVSREI-PV-BLICAE // PAX // R | | M / COMOB

2804var5 (2s: 2o/1r) DNANTHEMI-VSPFAVG

SALVSREI-PV-BLICAE // PAX // R | • | M / COMOB

Total (6s: 6o, 5r)

CONSULAR ISSUE- 1 JAN 468

Obverse: Pearl diadem bust facing, wearing consular robes, holding mappa in right hand and cruciform sceptre in left

Reverse: Leo and Anthemius, nimbate, enthroned facing in consular robes; each holds mappa and cruciform sceptre.

2806 (1s: 1o/1r) DNANTHEMI-VSPERPETAVC

VOTIS MVLTIS // RM / COMOB

SERIES 4

Obverse: Helmeted bust facing, pearl-diademed, cuirassed, in right hand spear over shoulder behind head, on left arm decorated shield (KFa3)

Reverse: Leo and Anthemius, facing, clad as cloaked soldiers (usually no cloaks trailing behind as in previous series), each holding spear in outside hand and together supporting globus cruciger between them.

2803 (1s: 1o/1r) DNANTHEMI-VSPFAVG

SALVSREI-PV-BLICAE // RM / COMOB

2807 (1s: 1o/1r) DNANTHEM-IVSPERPAVC

SALVSREI-PV-BLICAE // R | M | / COMOB

2807var (1s: 1o/1r) DNANTHEMI-VSPERPAVC

SALVSREI-PV-BLICAE // RM / COMOB

2808 (3s: 2o/3r) DNANTHEMI-VSPERPETAVC

SALVSREI-PV-BLICAE // RM / COMOB

2809 (2s: 1o*/1r) DNANTHEM-IVSPEAVC

SALVSREI-PV-BLICAE // RM / COMOB

2810 (2s: 1o*/1r) DNANTHEM-IVSPEAVC

SALVSRE-IPV-BLICAE // RM / COMOB

2811 (2s: 2o/2r) DNANTHEMI-VSPFAVC

SALVSRE-IPV-BLICAE // RM / COMOB

2812 (1s: 1o/1r) DNANTHEM-IVSPFAVC

SALVSRE-IPV-BLICAE // RM / COMOB

2813 (3s: 3o/2r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE // RM / COMOB

2814 (2s: 2o*/2r) DNANTHE-MIVSPFAVC

SALVSRE-IPV-BLICAE //  / COMOB

2814fn (1s: 1o/1r) DNANTHE-MIVSPFAVC

SALVSRE-IPV-BLICAE //  / COMOB

2814var (1s: 1o*/1r) DNANTHE-MIVSPFAVC

SALVSRE-IPV-BLICAE //  / COMOB

2815 (1s: 1o/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / COMOB

2816 (8s: 7o*/6r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / COMOB

2816fn (2s: 2o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE // RM >  / COMOB

2816var (1s: 1o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE // / COMOB

2817 DNANTHE-MIVSPFAVC }Requires

SALVSR-EIP-VBLICAE // / COMOB }confirmation

2818 DNANTHEM-IVSPFAVC

SALVSRE-IPV-BLICAE // RM / COMOB

2819 (1s: 1o/1r)  DNANTHE-MIVSPFAVC

SALVSRE-IPV-BLICAE // RM / COMOB

2820 (2s: 2o*/1r) DNANTHE-MIVSPFAVC

SALVSRE-IPV-BLICAE //  / CORMOB

2821 DNANTHE-MIVSPFAVC } Requires

SALVSR-EIP-VBLICAE // / CORMOB } confirmation

2821fn (2s: 2o/2r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / CORMOB

2822 (4s: 4o/1r) DNANTHE-MIVSPFAVC

SALVSRE-IPV-BLICAE //  / CORMOB

2823 (16s: 9o*/7r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / CORMOB

2823fn (3s: 2o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / CORMOB

2823var (1s: 1o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / CORMOB

2824 (3s: 3o*/1r) DNANTHE-MIVSPFAVC

SALVSRE-IPV-BLICAE // / CORMOB

2825 (6s: 5o/4r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / CORMOB

2825var (2s: 2o/1r) DNANTHEM-IVSPFAVC

SALVSR-EIP-VBLICAE //  / CORMOB

2826 (1s: 1o/1r) DNANTHEMI-VSPFAVG

SALVSR-EIP-VBLICAE // / CORMOB

Total (73s: 49o, 46r), one reverse shared with series five

NOTES:

2817: Kent lists the location of this specimen as Oxford, but it does not exist there. It may reside in Oderzo or Oslo, though I suspect that Kent may have confused  and  in his notes as he does in the preface (p. 196).

2821: Although Kent lists this specimen without a pellet above the chrismon, the plate photo shows a specimen with a pellet (which type is mentioned in the footnote). I do not doubt that such a specimen once existed, but I have not as yet seen one and I am uncertain that Kent did either.

SERIES 5

Obverse: Helmeted bust facing, pearl-diademed, draped and cuirassed, in right hand spear over shoulder behind head, shield slung behind left shoulder (MF3)

Reverse: Leo and Anthemius, facing, clad as cloaked soldiers (usually no cloaks trailing behind), each holding spear in outside hand and together supporting globus cruciger between them.

2830 (1s: 1o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / COMOB

2831 (16s: 8o*/6r*) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / COMOB

2831var (6s: 4o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / COMOB

2832 (2s: 1o*/1r*) DNANTHE-MIVSPFAVC (KFa3 series 4 bust)

SALVSR-EIP-VBLICAE //  / COMOB

2833 (6s: 3o*/4r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / •COMOB•

2833var (1s: 1o/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / •COMOB

2834 (1s: 1o/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / •COMOB•

2834fn (1s: 1o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE //  / xCOMOBx

2835 (14s: 7o*/4r)  DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE // :+: //  / COMOB

2835var (4s: 2o*/1r) DNANTHE-MIVSPFAVC

SALVSR-EIP-VBLICAE // :+: //  / COMOB

Total (52s: 19o,19 r), , one reverse shared with series four

GRAND TOTAL: 134 specimens , 77 obverses, 72 reverses

General Introduction Introduction Ravenna Milan

EUPHEMIA

SERIES 3

Obverse: Pearl-diademed draped bust with braided hair facing, wearing double necklace and spiked headdress with cross rising from front jewel of diadem (FF3, not FC3 as in RIC)

Reverse: Verina and Euphemia nimbate and wearing earrings, standing facing, each holding cruciform sceptre in right hand.

2805 (1s: 1o/1r) DNEVFYMI-APFAVG

GLORIAREI-PV-BLICAE / R | M | / COMOB

SERIES 4

Obverse: Pearl-diademed draped bust facing right wearing necklace, earrings, and headdress (F3).

Reverse: Victory standing left supporting a long jewelled cross.

2827 (9s: 1o/1r*) DNAELMARCEVFEMIAEPFAVC

VICTORI-AAVCCC / COMOB

2827var (4s: 1o*/1r*) DNAELMARCEVEEMIAEPPAVG

VICTORI-AAVCCC / COMOB

2828 (1s: 1o/1r) DNAELMARC-EVFIMIAEAVG

VICTORI-AAVCCC / COMOB

2829 (2s: 1o*/1r) DNAELMARCEVEEMIAEPPAVG

VICTORI-AAVCCC / CORMOB

GRAND TOTAL: 17 SPECIMENS, 4 obverses, 4 reverses

GRAND TOTAL FOR ROME: 151 SPECIMENS, 81 obverses, 76 reverses (so far)

General Introduction Introduction Ravenna Milan

PART TWO: LINKS

The links listed here do not include links within a single RIC catalogue number (e.g. 2823 to 2823), only those linking different numbers, and thus only those that will help with the analysis of the production order, found below.

KEY: (2s: 2o, 1r) = two specimens, two obverse dies, and one reverse die

--R-- = reverse link

            --O-- = obverse link

            <=> = link between different reverse types

ANTHEMIUS

SERIES 1-2

2801--R--2802 (2 s: 2 o, 1 r)

SERIES 3

2804var1--O--2884 (MD) (3 s: 1 o, 2 r)

SERIES 4

2809--O--2810 (PV < => IPV) (4 s: 1 o, 2 r)

2810/2812--O--2818 (IPV // RM / COMOB < => IPV // RM / COMOB(see Depeyrot, p. 158)

2816fn--O--2823 (RM >  / COMOB < => / CORMOB) (3 s: 2 o, 2 r)

2816--O--2823 ( / COMOB < => / CORMOB) (2 s: 1 o, 2 r))

Obv Rev RIC no. of Rev

1 (1 s) 1 (1 s) 1 = 2823fn EIP //  / CORMOB

2 (7 s) 1 (2 s) + 2 (1 s) + 3 (1 s) + 4 (1 s) + 5 (1 s) + 6 (1 s) 2 = 2816var EIP // / COMOB

3 (1 s) 4 (1 s) 3 = 2814var IPV //  / COMOB

4 (1 s) 6 (1 s) 4 = 2816 EIP //  / COMOB

5 = 2814 IPV //  / COMOB

6 = 2820 IPV //  / CORMOB

2824--O--2823var (IPV // / CORMOB < => EIP //  / CORMOB) (4 s: 3 o, 2 r)

SERIES 4-5

Obv Rev RIC no. of Rev

KFa3 EIP

1 (2 s) 1 (1 s) + 2 (1 s) 1 = 2823 = / CORMOB

2 (3 s) 2 (1 s) + 3 (2 s) 2 = 2823

3 (1 s) 3 (1 s) 3 = 2823

4 (3 s) 3 (1 s) + 4 (1 s) + 5 (1 s) 4 = 2823

5 (1 s) 3 (1 s) 5 = 2816

6 (6 s) 3 (4 s) + 6 (2 s) 6 = 2832/2831 =  / COMOB

MF3 EIP

7 (10 s) 6 (6 s) + 7 (1 s) + 8 (1 s) + 9 (2 s) 7 = 2831

8 (2 s) 8 (2 s) 8 = 2835 = :+: /  /COMOB

9 (3 s) 8 (2 s) + 10 (1 s) 9 = 2831

10 (5 s) 8 (3 s) + 9 (1 s) + 10 (1 s) 10 = 3831var =  / COMOB

11 (1 s) 9 (1 s)

12 (4 s) 9 (1 s) + 10 (3 s)

13 (3 s) 9 (2 s) + 10 (1 s)

Total: 44 s

2830--O--2833--O--2834fn (6 s: 2 o, 4 r) ( / COMOB < => / •COMOB• < => / xCOMOBx)

2835--O--2835var (6 s: 2 o, 2 r) ( < => )

EUFEMIA

2829--O--2827var--R--2827

Obv Rev Obv Legend Rev

1(6 s) 1 (2 s) + 2 (4 s) EVEEMIAEPPAVG 1 = CORMOB

2 (9 s) 2 (9 s) EVFEMIAEPFAVC 2 = COMOB

General Introduction Introduction Ravenna Milan

PART THREE:

THE ORDER OF PRODUCTION

The varieties of mint marks, symbols, and legend breaks on the reverses of the solidi of Anthemius are surprisingly complex and difficult to unravel. In RIC, Kent made no real attempt to make sense of any of it and enumerated the types simply with an eye toward ease of identification and cataloguing rather than actual mint output, listing the coins according to, first, COMOB, COMOB, and then CORMOB, then according to mint mark or symbol, reverse legend break, and finally obverse legend. This is exactly the right way to arrange it to facilitate the identification of a photo or coin in hand, but more should have been done to explicate the historical situation. The only method that would really have allowed him to make any sense at all of the detailed chronology was die linking, but he made no real attempt to do this, noting links only between 2801 and 2802, 2804 and 2884, and 2831 and 2832. Furthermore, he failed to systematically look for the appearance of engraved pellets over the various symbols in the field between the two imperial figures and seems to have missed a number of other variants of obverse and reverse (I have already discovered nine new variants for Rome that Kent missed and expect to find many more as my search expands). Nevertheless, it must be said that Kent’s general overall ordering of the issues, what I call series one to five, is correct and his achievement is best appreciated when compared with the much less successful attempts by Lacam and Depeyrot. It is my purpose in this part to determine the chronology of these multifarious and confusing types.

SERIES ONE TO THREE

The detailed arguments for the position and order of series one to three appears in the introduction. It is enough here to comment on chronological evidence that relates solely to Rome. The issues from Rome are the earliest of the reign and begin with the standard D4 type obverse portrait (series one) and progress from a single die pair with this obverse to a single KFa obverse die employed with the same reverse die (series two). This linked pair must begin the minting sequence. The next important clues for the chronology come from 2804 (series three) and 2806 (consular). 2804 begins with a reused die from Milan (= 2884) and a reverse cut by a Milan engraver as well (the mint mark does not appear to have been recut but the typical Milan PAS does appear). The rest of the obverse busts of 2804 are clearly the product of a Milan engraver, though the legends (and the PAX on the reverse) are strictly local and Roman. The engravers of the reverses are harder to judge, though. The style is better than the contemporary Milan reverses, but inferior to that of the local Roman engravers. In addition, the figures’ armour is different from that seen on earlier and later Roman reverses (both local and new style). It would seem, then, that Rome seconded the best engraver from Milan for these reverses. Once series three finishes, however, the Milan engravers disappear, never to be seen again in Rome. The obverse of the consular type, which must be dated to 1 January 468, is the product of new engravers brought from Constantinople. They clearly began training the local engravers but this effort seems to have been abandoned and for the most part these new workers did most of the rest of the engraving in Rome themselves. The reverse of 2806, however, was left in the hands of the locals. As a result all Eastern-style engraving must post-date 1 January 468 and series one to three must predate January 468.

SERIES FOUR A

The dies of 2803 are in a much finer style than 2802 and thus would seem to show influence from the new Constantinopolitan engravers, but both obverse and reverse are of a unique style that is not seen elsewhere at any mint and so whatever was going on with 2803 it was a failed experiment. The reverse includes the mintmark RM that is missing from 2801-2, but present on 2804. It also gives Anthemius a long spear that reaches into the legend (between the P and F), which for some unknown reason is a characteristic of series three at all three mints (see introduction). For these reasons, it must therefore be classified as the first of the new series four, rather than as the last of Rome’s series two.

The engraver of 2807 is one of the new Eastern-engravers and the same engraver of 2807var. The obverses of both share the form PERP in the place of the earlier PF (imported from the Eastern obverse legends of Leo by the new engravers), though they have a different legend break (2807: THEM-IVS; 2807var: THEMI-VS), that of 2707var being the basic break for 2802 to 2804, 2808, and 2811, while that of 2807 being the same as 2809-10 and 2812. More important, like 2803, they still share the odd series three characteristic of all mints that has Anthemius’ spear protruding into the legend: in 2707 it lies between the S and P, on 2807var between the P and E. No later die shares this characteristic. This guarantees that these two issues belong immediately after 2803 in the regular sequence. Further confirming this position is the fact that the reverse of 2807 was engraved by the same engraver who did the reverses of 2801-2 (and perhaps 2803), whereas all later reverses were completed by the new engravers. The design of the reverse is very similar as well. Unlike the reverses of 2807var and all the following types the two imperial military figures of 2807 stand straight at attention facing directly forwards. On the later reverses, the figures face slightly inwards and stand with their weight on the outer leg, their outer feet pointed outwards, and their inner legs bent slightly at the knee so that the inner feet face forward and down into the exergue. This gives the figures a more graceful and lifelike pose and creates more open space in the centre for the mint mark, instead of the box-sided rectangle of 2801-3 and 2807. This lack of space on 2707 forced the engraver to place the R between the legend and the left figure and the M in the centre. In 2803 the entire RM is placed between the figures but it is very small and cramped. After 2807 the RM returns to the centre field, where it remains for the rest of the first part of series four, down to 2813.

The appearance of PERP and PERPET on 2707 and 2808 also link these issues very closely with 2806, the consular issue, which has PERPET in imitation of the contemporary solidi of Leo. The characteristics noted above demonstrate that the PERP form must come first, and it is the main obverse legend break and titles that determine the order of the next issues: THEMI-VS = 2802 to 2804 (PF), 2806 (PERPET), 2807var (PERP), 2808 (PERPET), and 2811 (PF); THEM-IVS = 2807 (PERP), 2809-10 (PE), and 2812 (PF); and THE-MIVS = 2801, 28013 et seqq (PF). I suspect, therefore, that 2803, 2807, and 2807var precede the consular issue (PERP was perhaps used since the engravers could not be sure at first that PERPET would fit within Anthemius’ fuller legend), and that 2808 is contemporary with the consular issue. This will have the effect of moving the 2807 series from January 468 to December 467 with 2803. The PE obverses of 2809 and 2810 come next (a single obverse die linked to two reverses), followed by the return to the normal PF. Once the PF settles in, the legend break makes one more change, shifting back to the very first break of 2801, THE-MIVS, with 2813. This remains the normal break for the rest of series four and five, which shows that 2813 must come last in this sequence with RM in the reverse field.

The reverse legend break of 2801 to 2809 is PV. It becomes IPV for 2810-12, which must therefore belong together. The obverse die link between 2809 and 2810 confirms the timing of the change from PV to IPV. This leaves 2813 as the only RM / COMOB reverse with the next legend break, EIP. Kent notes the reverse die that shows a  (which I shall call a chi-rho, even though the loop is missing from all dies but one, 2816fn) engraved over an RM on an EIP / COMOB reverse, indicating that the latter preceded the former (thus a transition from 2813 to 2816). This provides further confirmation that EIP // RM (2813) must be the last of the RM / COMOB types since all earlier legend breaks with RM are PV or IPV. However, the remainder of the series four reverse types show both IPV and EIP, so although in four A there appears to be a progression from PV to IPV to EIP, EIP and IPV are in actuality contemporary types, and this must be kept in mind for the discussion below.

The order of the earliest issues so far discussed must therefore be 2801, 2802, 2804, 2803, 2807, 2807var, 2806, 2808, 2809, 2810, 2811, 2812, 2813 (series 4A), and 2816 (series 4B).

However, if changing obverse legend breaks coupled with an IPV reverse legend break and the RM mint mark belong in series four A, then 2818 and 2819 must belong here as well, since they both have IPV / RM reverses and the legend breaks THEM-IVS (2818; = 2807, 2809-10, 2812) and THE-MIVS (2819; = 2813 et seqq.). However, each has COMOB followed by  a six pointed star (not a chi-rho, as Kent indicates). Depeyrot claims that an obverse die used to strike an example of 2812 (or 2810: he misses the difference between PE and PF; see the concordance below) is linked to 2818 (p. 158; 2818 is in Vienna and I have not seen it yet). The conclusion can only be that 2818-19 form a linked issue parallel to 2810-12 and 2813, the two other types with which they share the most in common. The appearance of the star in the exergue is important here because it immediately precedes the appearance of the other reverse field symbols and, along with the double legend split of IPV and EIP, must indicate the first attempts by the engravers to mark the reverses of these solidi.

Although 2826 (EIP // / CORMOB) is listed last in series four by Kent, the legend break is THEMI-VS (= 2802-4, 2806, 2807var, 2808, 2811), and the engraver of the obverse legend is the same engraver who completed 2801-2, 2804var2-5, 2803, and 2805, but is never seen again on the obverses of Anthemius. The local Roman engravers always completed the G of AVG with a flourish, a tail that varied in size depending on the amount of space available for it (2801-4, Anthemius; 2805, 2808, 2829, 2827var, Euphemia). This goes back to the Roman reverses of the reign of Valentinian III and to the obverses and reverses produced at Ravenna and then Arles (with Roman engravers) for Leo and Majorian and then at Ravenna and finally Rome again for Severus. But the new engravers always represent the G as a C. Of series four and five only 2826 shows the characteristic flourished tail of the G seen in series one to three and on the obverses of all but the last of Euphemia’s obverse dies. Furthermore, the portrait is not the work of the bulk of the post-1 Jan 468 engravers, but closely mirrors the work of the engraver of the specimen in the RIC plate for 2811, DO 915 (2812), and Lacam 64 (2813). Similarly, 2825var bears the legend break THEM-IVS (= 2807, 2809-10, 2812, 2818), yet apart from these two groups no other obverse legend after 2812 displays any break other than THE-MIVS. 2824 and 2825 share their reverse type with 2826 and 2825var (EIP //  / CORMOB), a type that clearly harkens back to the original reverse at Rome ( / COMOB), the appearance of which so late in series four (in fact, so it would seem, right at its conclusion) would be most unusual. As noted above, EIP appears for the first time at 2813. Like 2825-6 the reverse of 2824 also has  / CORMOB but with the legend break IPV (= 2810-12, 2818-19). The reverse figures of 2825, 2825var, and 2826 each sport an obvious belly button, which is a characteristic of series four A (from 2808), but which disappears between 2813 and 2814 (though they do appear infrequently on other dies, such as 2823var). These belly buttons are, however, missing from 2824. The parallelism noted above in part two between the COMOB and CORMOB types and the COMOB / CORMOB links between 2824 and 2823var, 2816fn and 2823, and 2816 and 2823 (in two different places) in series four B demonstrate conclusively that the dichotomy between these two exergual markers was contemporaneous and not the result of a single chronological shift from one to the other. As a result there can be no objections to the two appearing together here in series four A. Finally, the  reverse type is the only field symbol that does not appear pelleted. In this way it sits outside the usual pattern of series four B symbols (see table below). As noted above in part two, there is an obverse link between 2824 and 2823var. Both have a THE-MIVS obverse break, while the reverse of 2824 is IPV //  / CORMOB and that of 2823var is EIP //  / CORMOB. 2823var is a specimen in the Ashmolean in Oxford that uniquely shows the hooked top of the rho in the chi-rho. All later examples of this symbol represent it as a kind of six-armed star. This would seem therefore to be one of the first representations of this symbol in the field, which for some reason was later simplified. If so, then  / CORMOB must precede it.

These facts indicate that 2824-2826 belong not at the end of the series four minting sequence (4B), but at the beginning (4A). Thus, the  / CORMOB was intended as a return to the original  / COMOB (2801-2), but with the mint mark that had displaced the star in 2807-2813 now incorporated into the exergue. This type would thus have run parallel to the RM / COMOB types at the beginning of series four, the RM issues with the break IPV (2810-12) and the CORMOB issues with the break EIP (2825, 2825var, 2816), though at the end EIP came over to the COMOB type (2813) and IPV came over to the CORMOB type (2824). For some reason, however, once started both sets of exergue markers and both legend breaks continued in use throughout series four, and it was only with the return of the ligatured RM at the very end of the series (2823) that CORMOB and IPV were finally dropped. The six-pointed star of 2818-19 did not continue but other symbols began to appear in the field instead.

The sequence of these four A types would therefore have been as follows.

TABLE 1: SERIES FOUR A PRODUCTION ORDER

I	THEMI-VS/PF THEM-IVS/PERP THEMI-VS/PERP THEMI-VS/PERPET THEM-IVS/PE	THEMI-VS/PERPET
	PV COMOB	VOTIS MVLTIS COMOB
	RM 2803, 07, 08, 09	RM 2806
II	THEM-IVS/PE THEMI-VS/PF THEM-IVS/PF	THEMI-VS/PF THEM-IVS/PF		THEM-IVS/PF
	IPV COMOB	EIP CORMOB		IPV   COMOB
	RM 2810-12	2826, 2825var		RM 2818
III	THE-MIVS/PF	THE-MIVS/PF	THE-MIVS/PF	THE-MIVS/PF
	EIP COMOB	IPV CORMOB	EIP CORMOB	IPV COMOB
	RM 2813	2824	2825	RM 2819

SERIES FOUR B

The links among 2823fn, 2816var, 2814var, 2816, 2814, and 2820 show not only just how complicated the symbol, exergue, and legend break variations were, but in particular that these variants could have had nothing to do with minting officinae, since EIP and IPV, pelleted and unpelleted, chi-rho and chrismon, and COMOB and CORMOB are all mixed in together in a single sequence that involved four obverse dies and six reverse die operating at the same time, with one obverse being particularly well represented (seven of the ten specimens). This does not mean that the marks do not refer to engraving officinae, but there an insufficient differentiation between or among engravers’ styles to be certain of this at this stage of research. It should be noted that the order of these reverses in the just-mentioned sequence is determined by the breakdown of the major obverse die, which develops a flaw and then a cud at AVC and then at the P right at the end of the sequence.

The evidence of the links and the number of surviving specimens indicates that chi-rho reverses (either pelleted or not) with either COMOB or CORMOB and for the most part the EIP legend break were the standard types throughout series four B, with the chrismon reverses (pelleted and not) interspersed among them. This is particularly true of 2816 and 2823, which show a number of links to one another. The most interesting is the sequence involving three 2823 reverse dies in which a single one of those dies is linked to five different obverse dies, of which one is further linked to two other reverse dies, one 2823, the other 2816. Neither of the latter reverses dies (at least as of yet) links with any of the other obverses.

In the end, it is clear from the evidence that for each exergual form in series four A and B, COMOB and CORMOB, there exist three inter-linked sets of marks: RM (COMOB) and  (CORMOB), which come first in four A; then , both with a pellet above and without; and (which I shall call a chrismon), again, both with a pellet above and without (see table below) in four B. The chrismon types are secondary to the chi-rho types. If the reverse legend breaks are taken into consideration, only EIP // CORMOB has both the pelleted and unpelleted symbols, and although EIP // COMOB has three of the four symbols, I suspect that an unpelleted chrismon, a variant of the unique pelleted 2815, will eventually appear. IPV // CORMOB has the two symbols unpelleted and IPV // COMOB has only a pelleted and unpelleted chrismon. The collection of further specimens will no doubt refine the present pattern. These eleven series four B combinations appear in no obvious pattern, though the bulk of the actual dies produced was made up of EIP //  / COMOB and CORMOB types, as even my very preliminary die counts demonstrate (lumping pelleted and unpelleted types together): EIP COMOB:  (2816) = 8 dies;  (2815) = 1 die; IPV COMOB:  (2814) = 4 dies; EIP CORMOB:  (2823) = 9 dies; (2821) = 2 dies (+1?); IPV CORMOB:  (2822) = 1 die;  (2820) = 1 die.

TABLE 2: SERIES FOUR B PRODUCTION

EIP COMOB	EIP CORMOB	IPV COMOB	IPV CORMOB
2816 2816var 2815	2823 2823fn 2821 2821fn	2814fn/var 2814	2822 2820

SERIES FIVE

The complex linking sequence detailed above in part two shows that series four ended with 2823 (EIP //  / CORMOB) and one 2816 (EIP //  / COMOB), and that this was then replaced with a much less complex but no less puzzling system that was based upon EIP // /COMOB (2832-1), much as series four had been based on EIP //  / CO(R)MOB. It also shows that just as was the case in series four, where the sequence of 2814, 2814var, 2816, 2816var, 2820, and 2823fn shows that all four pairs of variants—chrismon and chi-rho, IPV and EIP, pelleted and unpelleted, and COMOB and CORMOB—were engraved at the same time and supplied apparently randomly to the anvil as they were completed, reverses of series five with varying symbols and marks also apparently randomly interrupt the main sequence. The linking sequence involving 2835, 2831, and 2831var, in which seven obverse dies appear to have been used simultaneously or nearly so with three reverse dies, reveals a complexity that would have been unnecessary had different minting officinae been in use. It would be interesting to know how such a large and closely linked group of coins survived.

Thus 2832, 2831, 2831var, and 2835 are very closely linked, as are 2830, 2833, and 2834fn. The only remaining type is 2834. It bears •COMOB• like 2833 and probably belongs close to it in the sequence.

EUPHEMIA

Kent was correct in his placing of 2805 (pp. 194-5). It matches very nicely the style of portraits and legends of 2804, produced by seconded Milan engravers. The reverse design of two standing imperial women is based on an unpublished consular solidus of Valentinian III issued on 1 January 426 from Ravenna, which depicts Theodosius and Valentinain standing facing each holding a mappa and cruciform sceptre (image). It in turn seems to be based, like the MF bust type of Anthemius, on unknown imperial iocongraphy, since the only comparandum is a very small series of double solidi produced in January of 346 by Constantius II (image). Given this pedigree and the depiction of Leo and Anthemius jointly on the regular reverses of Anthemius, there can be no doubt that this reverse depicts Verina and Euphemia, as Kent correctly suggests (p. 195), not Euphemia and Alypia, as indicated by Grierson and Mays (pp. 260-1).

In his introduction Kent dismisses the idea that the next issues, 2827-9, belong with the star reverses to which he gives the numbers 2824-6 (‘[They] appear at first sight to belong to the  phase of this series’, p. 196) and concludes that they must belong with the starred reverses of 2801-2 (‘This group should therefore be placed with the Rome issue with the star at the beginning of the reign’, p. 196). In spite of this introduction, in the catalogue he places them immediately after the star reverses of 2824-6, at the very end of series four (2827-9). Kent thus appears to have been conflicted about the dating.

This is not altogether surprising since these issues present us with a problem. First of all we need to deal with 2828, which is slightly cruder than the fine style of 2827 and 2829. It has the same improved local style that is seen in 2803 and on the obverse of 2826 and the reverse of 2807, and it displays the local and therefore early tailed G of AVG. It thus belongs at the very beginning of series four A. The remaining three catalogue numbers, two obverse and two reverse dies with an obverse link, are more difficult. The sequence must start with 2829, both for the use of the tailed G on the obverse and the appearance of PP, instead of PF, which mirrors the use of PERP, PERPET, and PE on the coins of Anthemius (2806-10). The second obverse has a legend by the new engravers with PF and their typical C form of the G. The star in the reverse legend of 2827-9 with CORMOB or COMOB and no other mint mark also points to an early date, when the  / CORMOB reverses were being produced for Anthemius. This is the evidence that led Kent to suggest an early date (though he did not realize that 2824-2826 were early) and it is solid and consistent.

However, the appearance of ten of these coins (linked examples of 2827, 2827var, and 2829) in the Vestal hoard suggests that they were struck at about the same time as the bulk of the hoard, therefore at the very end of series four or even the beginning of series five: Vestal 18-19 (2823, KFa3) have an obverse link to Vestal 348-55 (2832, KFa3; both series four), which have a reverse link to Vestal 22-144 and 146-347 (2831, MF3, of series five). The other coins of the Vestal hoard from the mint in Rome are somewhat earlier, but they are very few in number: Vestal 15 = 2809, 16 = 2820, and 17 = 2825. Vestal 15 and 17 are therefore roughly contemporary and early (in fact exactly where the Euphemia types should be placed), while 2820 cannot be dated within series four, apart from the fact that it is neither very early nor very late. This would suggest that the issues of Euphemia found in the Vestal hoard are contemporaneous with the bulk of the Anthemius issues, thus late series four or early series five. The similarity of style and technique on these coins shows that no great time separated their engraving, so 2829 cannot be a reused die from earlier in the period while 2827 is a later work.

However, we have no proof that what we have in the Vestal hoard is in fact a single payment. Vestal 1 is a vota solidus of Constantius II dating to the late 340s (RIC 8 Nic 32). The hoard also contains seven solidi of Valentinian from Ravenna and Rome from both early (Vestal 5 = RIC 2010) and late in the reign (Vestal 4 = 2024), two solidi of Libius Severus (9-10), eight Constantinopolitan solidi of Marcian (366-73), and twenty-four solidi of Leo (374-97), also from Constantinople. Certainly the major block of the hoard (18-19, 22-355) came almost directly from the mint, but the rest of the collection process is unknown to us. Like the solidi of Valentinian the solidi of Euphemia could have been collected earlier, at the same time as nos. 15 and 17, and added to the hoard when it was dumped in the Vestals’ toilet.

As a result I believe that Euphemia’s solidi belong with the first series of  / CORMOB and RM / COMOB issues in period II of series four A.

REVISED ORDERING OF THE SOLIDI FROM THE MINT AT ROME

Series 1

2801

Series 2

2802

Series 3

2804, 2805

Series 4 A

I

2803, 2828

2807, 2807var

2806

2808

2809

II

2810, 2811, 2812; 2826, 2825var; 2818

2829, 2827var, 2827

III

2813, 2824, 2825, 2819

Series 4 B

2816 (all), 2815, 2814 (all), 2823 (all), 2821 (all), 2822, 2820

2823

2832

Series 5

2831

2831, 2835, 2831var, 2835var

2830, 2833, 2834fn

2834

General Introduction Introduction Ravenna Milan

CONCORDANCES

D = Georges Depeyrot, Les Monnaies d’or de Constantin II à Zenon (337-491) (Wetteren, 1996), pp. 157-60 and plate 10-11.

DO = Philip Grierson and Melinda Mays, Catalogue of Late Roman Coins in the Dumbarton Oaks Collection and in the Whittemore Collection. From Arcadius and Honorius to the Accession of Anastasius (Washington, DC, 1992).

F = Joan M. Fagerlie, Late Roman and Byzantine Solidi Found in Sweden and Denmark (NNM 157; New York, 1967).

L = Guy Lacam, La fin de l’Empire romain et le monnayage or en Italie, 455-493, vol. I (Lucerne, 1993), pp. 445-63, 491 (plates CXI-CXX, CXXVII) and vol. 2, plates 24-7, 32.

UB = Oscar Ulrich-Bansa, Moneta Mediolanensis (352-498) (Venice, 1949).

V = Lucrezia Ungaro, ‘Il Ripostiglio della casa delle Vestali, Roma 1899’, Bollettino di Numismatica 4 (1985), 47-160.

2801 L 4, D 57/1

2802 DO 918 = L 92

2803 L 51

2804var1 DO 908 = L 31

2804var2

2804var3 UB Mr

2804var4 L 30

2804var5 L 29 = D 70/1

2805 DO 934, D 59/7

2806 D 57B/1

2807

2807var

2808 F 157, D 59/1

2809 V 15 = L 53, L 52, L 54

2810 F 158

2811 DO 915 = L 59

2812

2813 L 64, L 65, D 59/6

2814

2814fn

2814var DO 919 = L 58

2815 L 60

2816 DO 916 = L 63, L 61, V 20, F 163, L 62 = D 60/1

2816fn

2816var V 21

2817

2818

2819 D 58/1

2820 V 16 = UB Mp = L 70/69, D 62/1

2821

2821fn L 74

2822 F 164 = L 55

2823 L 73, UB Mp*, F 165, V 19, V 18, L 71, L 72, D 61/1

2823fn

2823var

2824 L 57, L 56

2825 V 17 = L 69/70, D 63/1

2825var

2826 L 75

2827 V 356, V 357, V 359, V 362 = L 148, V 364, L 151, L 149, DO 933 = L 150

2827var V 358, 360-1, 363

2828 L 153 = D 66/1

2829 V 365 = L 152, D 64/1

2830 L 50

2831 L 41, V 22-347, DO 925, L 44, L 46, L 37, F 162

2831var L 43, L 45, L 42

2832 V 348-55 = D 65/2

2833 L 48, L 49, F 160, D 67/1

2833var

2834 F 159

2834fn D 68/1

2835 L 40, L 33, L 34, L 35, DO 924 = L 39, L 38, L 36, D 69/1

2835var F 161

Georges Depeyrot, Les Monnaies d’or de Constantin II à Zenon (337-491) (Wetteren, 1996), pp. 157-60 and plate 10-11.

57/1 2801 (‘Peut-�tre une imitation’)

57B/1 2806

58/1 2819

58/2 2818

59/1 2808

59/2 2807var

59/3 2809 misdescribed

59/4 2809 misdescribed

59/5 2812

59/6 2813

59/7 2805

60/1 2816

61/1 2823

61/2 2822

62/1 2820

62/2 2821

63/1 2825

63/2 2826

63/3 requires confirmation (THEMI-VS with IPV //  / CORMOB; in Vienna)

63/4 2824

64/1 2829 reverse with 2827 obverse

65/1 2832

65/2 2831

66/1 2827 (but photo is 2828)

67/1 2833

68/1 2834fn

69/1 2835

70/1 2804

APPENDICES

NOTES ON THE VESTAL HOARD

The weights given for many of the specimens in the catalogue are suspect, especially those for nos. 348 to 355, which are in the range of 4.50s and 4.60s, reaching a high of 4.68 g. Lacam’s figures for the Vestal specimens he illustrates are almost always different and more normal (i.e. in the 4.40s).

The following are the correct die identities for the Vestal hoard:

Obverses: 18-19 and 348-55 (2823, 2832)

Obverses 22-347 (2831)

Obverses: 356-7, 359, 362, 364 (2827)

Obverses: 358, 360-1, 363, 365 (2827var, 2829)

Reverses: 18-19 (2823)

Reverses: 22-144, 146-355 (2832, 2831)

Reverses: 356-64 (2827)

A NOTE ON ULRICH-BANSA

Plate Mo is a mule of two different BM specimens (Bank Coll. R. 83 [2833] and 1867 1.1.975 [2826]).

NOTES ON LACAM

Lacam 4 (2801) is illustrated with the Ravenna specimens and 92 (2802) is illustrated with the Milan specimens. Lacam nos 45 and 46 (Museo Bottacin, Padova and BM 1860.3.29.269) do not appear in the main text, while Pl. CXIII no. 5 (DO 925) does not appear in the summary of numbered photos in vol. 2. In Pl. CXVI the obverse and reverse photos have been swapped (i.e. upper obverse belongs with lower reverse, and vice versa). In Pl. CXVII the solidus labelled ‘B.M.’ in the bottom row is DO 916.

General Introduction Introduction Ravenna Milan