How Were Roman Coins Made? Inside the Imperial Mint

Short answer: a Roman coin was struck, not cast. A disc of metal called a flan was placed on a fixed lower die, a second die was held on top of it, and a workman hit that upper die with a hammer. One blow, one coin. Everything else — the refining, the casting of blanks, the engraving of dies, the counting and weighing — existed to keep that single blow supplied.

It is worth pausing on what that means. There was no press, no reducing machine, no master hub. Every die was cut by hand into hardened bronze or iron by an engraver working at roughly life size on a surface the width of a thumbnail, and every coin that survives is the impression of one particular hand-cut object, hit once. That is why no two Roman coins are quite identical, and why the marks of the process are still legible on a coin you can hold today.

1. Getting the metal right

The mint began with bullion: tax receipts, war booty, recalled old coin, and mine output. This was melted in crucibles and brought to the standard required for the denomination being struck — refined for gold, alloyed for silver and for the brass and bronze denominations (Burnett, 1987).

The alloying was more deliberate than it looks. Sestertii and dupondii were struck in orichalcum, a golden brass of copper and zinc, while the as was struck in plain copper — a visible colour difference that let users tell denominations apart at a glance. Silver coinage was mixed to a target fineness that the state adjusted repeatedly over the centuries, and modern analysis of that alloy is precise enough to track policy changes issue by issue (Butcher and Ponting, 2014). The long decline in that figure is the subject of our piece on how the denarius lost its silver.

2. Making the blanks

The blank disc — the flan, sometimes called the planchet — was produced in quantity before any striking took place. Two methods are attested in the archaeology: casting individual discs in open or two-part moulds, sometimes in linked rows that were then snapped apart, and cutting discs from cast strip or from the end of a cast rod (Metcalf, 2012).

Weight control happened here, not later. Roman mints worked al marco — by the batch — meaning a fixed weight of metal was required to yield a fixed number of coins, with individual pieces allowed to vary around the average (Grierson, 1975). This is a crucial point for collectors: two genuine denarii of the same issue can differ by a few tenths of a gram without either being suspect. Only for gold was the tolerance tight, and heavy flans were sometimes filed down before striking, leaving parallel adjustment marks that survive under the design.

3. Cutting the dies

The dies were the mint's skilled work and its bottleneck. Each was a block of hardened metal with the design engraved into it in negative and in mirror image, cut with gravers and finished with small punches for repeated elements — the beaded border, the letters of the legend, the pupil of an eye (Vermeule, 1954).

The two dies were not equivalent. The lower die sat fixed in an anvil and took the design that mattered most, which under the empire meant the imperial portrait; because it was supported from beneath, it survived far longer. The upper die was held in the hand or in a tongs and absorbed the hammer blow directly, so it cracked, spread and failed first. This is why the reverse of a Roman coin is more often weakly struck or visibly damaged than the obverse, and why the portrait is usually the sharper of the two faces.

Because engravers worked individually and by hand, the same issue could be cut by several artists of very different ability. Style is therefore evidence: an engraver's characteristic treatment of hair, drapery or lettering can tie coins to a mint, a period, and sometimes to a single workshop — and a style that fits nothing known is one of the strongest signals in authentication.

4. The strike

The striking itself was fast and physical. A flan — often heated to soften it, particularly for the larger bronzes — was set on the lower die, the upper die was positioned over it, and a hammer blow drove the metal into both engravings at once.

Experimental replication of the process suggests a small team could produce coins at a rapid rate, and that die failure rather than labour was the limiting factor (Sellwood, 1963). Estimates for how many coins a die could yield range widely; the working consensus places a lower die in the order of tens of thousands of strikes, with the upper die failing well before that. Multiply those figures across the mints of the empire over six centuries and you arrive at the enormous surviving population that keeps common Roman coins so inexpensive today (Harl, 1996).

The dies were not locked in alignment. The rotational relationship between the two faces — the die axis, quoted as a clock position — was effectively random on many early imperial issues and became increasingly standardised at 12 o'clock or 6 o'clock in the later empire. It is a routine field in modern catalogue descriptions and a useful check when comparing a coin against a published example.

5. The workshop floor

A Roman mint was an industrial establishment with a divided workforce. Funerary inscriptions of the familia monetalis at Rome name the specialisms individually: melters who handled the crucibles, flan-makers, engravers who cut the dies, the suppostor who placed each blank on the anvil, and the malliator who swung the hammer (Burnett, 1987). Above them sat state officials — in the Republic the annually appointed board of moneyers whose names appear on Republican coinage, and under the empire a procurator answering to the imperial administration.

This was a workforce of real size and real leverage. In AD 271 the mint workers of Rome rose in revolt under the official Felicissimus, and the suppression of that revolt was violent enough to be remembered by later historians. Coin production was not a quiet clerical activity; it was a strategic state industry.

6. From one mint to a network

For much of the early empire, Rome and Lugdunum carried the bulk of production. The pressures of the third century and the administrative reforms of Diocletian turned that into a formal network of regional mints — Londinium, Treveri, Siscia, Thessalonica, Nicomedia, Antioch, Alexandria and others — each divided internally into officinae, or workshops, striking in parallel.

That structure is the reason late Roman coins carry the small letters and numerals in the exergue beneath the reverse design. They are not decoration: they identify the mint and the workshop within it, and they are frequently the fastest route to dating a coin. We cover how to read them in mint marks and officinae, and the separate senatorial marking of bronze in what SC means on a Roman coin.

What the process leaves on the coin

Nearly every physical oddity that worries a new collector is a by-product of hand striking. The table below separates the ones that are normal from the ones that are not.

That last row is the one to memorise. Because Roman coins were struck and struck coins have no mould line, a seam running around the edge, a pitted granular surface or a suspiciously soft rim indicates a cast — either a modern forgery or, less often, an ancient one. Understanding manufacture is not an academic exercise; it is the foundation of every practical authenticity check you will ever make.

Every Roman coin is a fossil of a single moment: a hand-cut die, a heated blank, and one hammer blow that will never be repeated in exactly the same way again.

Frequently asked questions

Were Roman coins struck or cast?

Struck. With the exception of the heavy cast bronze of the early Republic, and of ancient and modern forgeries, every Roman coin was made by placing a metal blank between two engraved dies and hitting the upper die with a hammer. Cast surfaces on a supposedly imperial coin — a grainy texture, a soft rim, a seam around the edge — are one of the clearest warnings that it is a copy.

How many coins could one Roman die produce?

Estimates vary widely and depend on metal, temperature and the die's own quality. Modern experimental striking and die-study analysis suggest a working range in the order of ten to thirty thousand coins from a lower die, with the upper die failing considerably sooner because it absorbed the hammer blow directly. Any single figure quoted with confidence should be treated with caution.

Why are Roman coins not perfectly round?

Because the blank was cast or cut individually, and because the striking blow spread the metal unevenly. Roman mints controlled the weight of a batch rather than the shape of each piece, so irregular outlines, off-centre designs and flat spots where the flan was too small are entirely normal on genuine coins.

Who actually worked in a Roman mint?

A large staff of slaves and freedmen under state supervision, organised by trade. Funerary inscriptions from Rome name melters, flan-makers, engravers, the man who positioned the blank on the die and the man who swung the hammer. The mint was an industrial workplace, and in AD 271 its workforce was numerous and organised enough to mount a revolt against the emperor Aurelian.

References

• Burnett, A. (1987) Coinage in the Roman World. London: Seaby.
• Butcher, K. and Ponting, M. (2014) The Metallurgy of Roman Silver Coinage: From the Reform of Nero to the Reform of Trajan. Cambridge: Cambridge University Press.
• Grierson, P. (1975) Numismatics. Oxford: Oxford University Press.
• Harl, K.W. (1996) Coinage in the Roman Economy, 300 B.C. to A.D. 700. Baltimore: Johns Hopkins University Press.
• Metcalf, W.E. (ed.) (2012) The Oxford Handbook of Greek and Roman Coinage. Oxford: Oxford University Press.
• Sellwood, D.G. (1963) 'Some experiments in Greek minting technique', Numismatic Chronicle, 7th series, 3, pp. 217–231.
• Vermeule, C.C. (1954) Some Notes on Ancient Dies and Coining Methods. London: Spink.

Referencing follows the Harvard (author–date) system. Where figures are given as approximate ranges, they are drawn from the works above; published analyses differ by mint, issue and method, and estimates of die output in particular remain contested.