In this neat thread on coins with a blank side (by @Ocatarinetabellatchitchix), I posted the following coin. Judging by the obverse, it is a Trier mint product, either RIC 520 or 539, with a missing GLORIA EXERCITVS reverse: I asked for opinions on what the odd nub was on the reverse. Normally these blank-sided coins result from two stacked blank flans being struck by accident, resulting in one coin with an obverse plus a blank side, and another with a reverse and a blank side. I've never seen another example with a nub, they're usually smooth with a crescent where the edge of the intervening blank flan was. (There is a similar crescent on my coin, but as you'll see below, I doubt this is an impression made by a blank.) I got a couple of interesting answers: I like this theory! I had wondered whether something of the sort might explain it, but hadn't looked into it at all. On the strength of @Orange Julius's and @TIF's thinking, today I did. And ended up putting enough work into it that I felt it should be moved to its own thread. First, a couple of points to note in favour of the theory: 1) If there were a simple linear nub and socket system for aligning the dies, as suggested by my coin, then there would be two possible ways to orient the obverse and reverse dies with respect to each other, 180° apart. And this is exactly what we find: these Trier GLORIA EXERCITVS issues always have a 12h or 6h die axis +/- 1. (This seems to apply to the GLORIA EXERCITVS issues more generally.) Also, given the enormous volume of production, we would expect some automatic way of keeping the die axis within such a narrow range. 2) According to "The production of ancient coins" by Wickens which purports to summarize our knowledge of ancient coin production, "Some [Roman] dies may have had pegs or notches to help guide the minters in orienting the punch." I'm not sure where he got this information, but it seems the idea has already been proposed, presumably with some support. (That evidence might just be the consistent die alignment I note in (1).) In a few places (for example The Art Institute of Chicago), I found the hypothesis that "the image could be engraved directly into the base of a cylindrical or pyramidal piece of metal or, more commonly, into a bronze disk that was fitted into an iron punch or collar." The disk/collar method would allow for an iron striking mechanism to have new dies easily inserted as needed. The alignment socket would be in the iron base, and the nub in the die disk. Now take a look at the following doctored image of my coin. This is as though the metal is partially transparent, and you're looking directly at the reverse, while seeing a ghostly image of the obverse through the coin: Ignore the blue circle for the moment, and notice this: the edge of the obverse die lines up perfectly with the crescent line on the reverse. That's an odd coincidence if the crescent on the reverse were caused by a blank flan. A much more probable explanation is that the crescent on the reverse is just a product of the obverse die strike. (Imagine doing this to a piece of Play-Doh - the disk would get crushed on both sides along the edge of the die.) The blue circle indicates the position of the obverse die. Another interesting point to note is that the obverse is struck off-centre. Now, the production quality of these Constantinian AE coins was amazingly high, and off-centre strikes are highly unusual. Take a look at the results from this acsearch - almost every coin is perfectly centred, give or take a tiny bit! Given the volume of production, this is pretty amazing. This suggests to me that, just there was some automatic method to align the dies (the nub/socket), there was also an automatic method to centre the coins. The easiest way to do this is to make sure the anvil die has a depression which is a good fit for the flans - though a bit larger to accommodate varying flan shapes. The result would be almost perfectly centred coins, with a little wiggle room - just as we see. Putting this all together, here is a depiction of a die setup that would explain both the consistent production aspects, as well as my odd coin: In the left cross-section, we see everything operating as it should. There's an anvil which has a depression to accept the replaceable die disk. At the bottom of the die disk is the nub which keeps the anvil die aligned, matching with a socket in the anvil. The flan (green) sits in a depression within the die-disk that ensures proper centring. In the right cross section, the replaceable die disk is missing, with the result that 1) the die-alignment socket in the anvil is impressed into the flan, and 2) the flan is not properly centred, it's free to move within the anvil's depression, and so 3) the hammer die creates matching crescents on both obverse and reverse. Producing my coin, in other words! However... it's not all smooth sailing for the theory, I'm afraid. Here is one significant worry: As TIF notes, on my coin we would expect the impression of the alignment socket to be from the hammer die, given that the portrait is normally on the anvil die. That's not how I've drawn it - instead, my setup involves the portrait being on the hammer die. Now it's of course possible (and even likely) that the hammer die had a replaceable die disk as well, but I'm having trouble imagining a setup whereby that produces my coin. In particular, it seems hard to explain the off-centre strike. If the anvil die disk were present, and it was the hammer die disk that was missing (exposing a socket), wouldn't the flan still be centred properly within the anvil die disk depression? It's hard to imagine a centring mechanism that involved the hammer die rather than the anvil die. (Here I'm assuming the two weird aspects of my coin, the nibbed blank side and the off-centre strike, are somehow connected. That assumption might be false! But we'd still need an explanation for both aspects.) One possible conclusion to draw might be that the portrait was indeed on the anvil die for producing these Constantinian coins. Another possible conclusion could be that TIF's improbable hypothesis is actually correct - my coin was produced in the normal way, with two blank flans, one of which just happened to have a weird rectangular divot in it. What do you think? And please post your family of Constantine AEs, especially the imperfectly struck ones!
As far as I know the reverse is always the hammer die, or at least that's the way Vagi explained it to me, and would be held with thongs while the obverse was anchored. I suppose if the opposite occurred and the obverse were struck instead, then it would show slightly concave, assuming you can see the concavity at all.
Hmm… if the hammer is the reverse, what if it was just that the die fell off the hammer before or during striking? I can imagine a Kevin, early after a weekend at the games, trying to just get his quota of struck coins for the day. He has a splitting headache and his hammer die keeps falling out, causing him to miss-strike a number of coins. He’s having a rough one so, just utters some unkind words at the malfunctioning minting tools and tosses the coins in the bag when normally he’d put them in the reject pile. That movie works in my mind, so that’s what I’m going with. Edit: what did these later GLORIA reverse dies look like? I’d assume, they’d somehow affix to the set up and be easily interchangeable if something broke. Edit 2: Went back and noticed that TIF said the same thing already… so we’re dreaming in the same direction.
Interesting theory. I've just looked through ~500 results on ACsearch for similar coins, and found only this one (but it's not Roman): https://www.acsearch.info/search.html?id=3059724 There were numerous LRB blanks, both obverse and reverse, but none with the bulb. I would like to see more examples. If dies where indeed shaped as hypothesized, one would easily find more example.
Good work, Sev, and terrific illustrations! I hadn't pictured the removable die being inset as you drew. I envisioned it being more of a butt joint with a mortise and tenon but thinking about it more, it's difficult to see how that would be a secure attachment capable of lasting through thousands of strikes. As @Roerbakmix noted, you'd expect more examples of coins like yours. Gotta go to work now . I'd rather keep thinking about coins and this dilemma! Tossing this in here because I came across it while looking for info to explain your coin. This is purportedly a hinged set of Roman coin dies. I'd like to find better images to see if there is engraving on the faces or if there were removable dies (which are gone). Don't have time to search but a preliminary image search turned up only this tiny image and one larger version that was a cut-out of the dies. It was larger but the resolution was still poor. Looks like a museum tag on the right side. Maybe a museum search would turn up this or more like it.
In the earlier Empire, the reverse was almost always the hammer die and that was what defined it as the reverse. In those days, brockages all showed an incuse of the obverse. We start seeing reverse brockages around the time of the Gallic Empire. https://www.forumancientcoins.com/dougsmith/feac72bro.html I suspect some of these coins were struck with hinged pincher dies. http://www.grunalmoneta.co.uk/history.html I do not know the origin of the above photo but there is a dark rectangle on the die face that could support Sev's theory (which seems quite reasonable to me). I have never seen a coin with this 'mechanism'. http://muzeydeneg.ru/eng/?p=665
S.A., I like your theory & believe it has weight. It would be great if you had access to photos from a hoard of coins like the recent Epfig Hoard, that contained many coins of AE3 size that were struck from the Trier Mint. Possibly another freak coin like yours might be in that hoard. The Trier Mint probably had the largest output of AE3 coins of all the Roman mints in that time period, & if your theory is right it must have been applied to all the coin types & all the officina from that mint. One advantage of looking at slabbed coins from the Trier Mint, you can get an accurate measure of die alignment from obverse to reverse. There certainly is a repetition of die alignment if you examine enough coins that speaks well for your theory. Pictured below are a few coins from my collection. Officina 1 Officina 1 Officina 2 Officina 2 Officina 2 Unlike the other 4 coins I posted, this coin has a near 12:00 o'clock alignment (medallic alignment), possibly because it's a consular issue ?
The point about concavity is a good one. I find it difficult to detect any concavity on my lrb's - the portrait side tends to create an illusion of slight concavity because of its large, raised central surface, but the illusion is dispelled if the reverse also features a portrait (like the Constantine/Sol bust coin). One coin I have which clearly has a concave obverse (thus suggestive of a hammer die) is this antoninianus of Maximianus from Lugdunum: And as @dougsmit pointed out, reverse brockages become much more common after 260 or so, suggesting that the location of the portrait may have at least sometimes been on the hammer die. (For those wondering why: It's a lot easier for a mint worker to notice and remove a stuck coin on the anvil die, so brockages are usually formed by coins stuck on the hammer die. A reverse brockage would then be expected to be produced by a previously struck coin having its portrait side wedged into the hammer die.) Reverse brockages are well known in Constantinian era coinage. Here's mine: And here's one of a GLORIA EXERCITVS from Trier! (Sold by Leu, not mine.)
I agree this is a possibility, for sure. Then the only thing left to explain would be why my coin is so off-centre. If the replaceable die disk fell out of the hammer, you'd still expect the coin to be well centred on the obverse die depression sitting on the anvil. Of course it could be two different accidents coinciding: the die-disk falling off the hammer at the same time as a misplaced flan. Less likely, but certainly possible! Then again, we'd expect more instances of a diskless hammer die striking its nub over a well struck obverse, then. (A low probability coincidence of two events suggests higher probabilities for each of the events occurring alone.) Unless... maybe the hinged dies that @TIF and @dougsmit showed could explain it? If they had replaceable die disks that were thick enough, we'd expect an off-centre strike when one of them was missing - the die missing its disk would travel too far and not line up properly. I actually thought about that possibility, but then we'd expect one side of the portrait to be struck up much more than the other, due to the angle of impact - it would cause an angular misalignment (see image below). And we don't see that on my coin, it looks more like a pure "offset misalignment" as in my diagram in the OP. Still... a possibility, I admit! (I don't think anybody knows exactly when the hinged dies were introduced.)
Great article. I agree that the removable disk was in fact removed, perhaps accidentally. And the coin makers struck a flan directly into the groove on the anvil which presumably had that alignment ridge. For some reason it got into the batch of "good" coins and was not re-heated and re-struck.
Thanks for looking into this - I would like to see more examples too! I agree that the theory would require more examples before it could be accepted. Maybe in the coming years someone will google this thread and post their example. I can think of a couple reasons for the error's rarity, if the theory is right: 1) If the missing die disk were on the anvil, as I've represented it, it would be really hard for a mint worker to fail to notice it was missing. An extra-Kevinish mistake! (Perhaps their rarity offers a little support for the idea applying to the anvil die only? Small, moveable hammer dies were generally easier to replace, so maybe there was a need for a replaceable die disk on the anvil only?) 2) The die-disks were surely anchored - maybe by iron dowel pins, for example, or a collar that could be tightened. Perhaps the anchoring system was simply extremely reliable resulting in very few errors of this type.
Yes...but: If the disks were just held in place by friction and a coin were struck like the one here, I could see it doing considerable damage to the die making repair to the hinged holder a big job. Perhaps more care was taken to avoid this than was, for example, to avoid simple clashes or brockages. I would fear that a solid attachment would make changing out the design inserts more complicated (slower) and I do believe we would all be shocked at how fast a pace was used making these coins. I doubt anyone has ever tried to replicate this system. Whenever you see modern mint recreations, we see people doing it like the medieval engravings show. What is the oldest illustration showing mint operation? This weekend at the coin show I was asked what kind of hammer I thought was used to strike these coins. I do not know BUT I suspect that sestertii and tetradrachms had a full size sledge while AE4, obols and medieval pennies may have been done differently. I even believe that a hammer with a curved head would be easier on the wrists and, whether it was your wrist or that of your slave, making technical advances as needed would be part of the job. I guarantee you one thing. We do not know all we would like to know about all this but we have learns things in the last 50 years. I do not see that changing. I only hope our great grandchildren have the opportunity to study these questions and come up with things beyond our dreams.
I still haven't been able to find more info on these hinged dies. But I do see that dark rectangle, @dougsmit - good catch! Another reason to suppose the error (as I envisage it) would be rare. Thanks! Thanks for posting these beauties, @Al Kowsky! When I researched the die alignments, I just used acsearch restricting to e.g. Roma or CNG, who always report the die alignment. It's always 6h +/-1, or 12h +/-1, with no obvious preference for coinage axis (6h) over medal axis (12h). So it's probably just a fluke that your consular issue is your only one struck with a medal axis. 100%!!
This is a great thread. I will have to circle back to it when I have some time to think on the question and look into it more. My initial reaction is that you are definitely onto something here @Severus Alexander . Nice work! In the meantime here is my GLORIA EXERCITVS Roman Empire Constantine II as Caesar (AD 317 - 337) AE Follis, Siscia mint, struck ca. AD 330 - 333 Obv.: CONSTANTINVS IVN NOB C. Laureate, cuirassed bust right. Rev.: GLOR-IA EXERC-ITVS. Two soldiers standing, facing each other with spear in outer hand, inner hand on shield resting on ground; two standards between. Ref.: RIC VII 220 (Siscia) Purchased in Jerusalem, Nov. 2016
The only specific reference to hinged dies I could find is this, from an article "ANCIENT METHODS OF COINING" by George F. Hill, from Numismatic Chronicle 1922 (available on JSTOR). He seems a bit confused taking about an aureus of Constans I. Maybe he meant an aureus of Constantius I, or a solidus of Constans I. I was hoping the Beaumont-sur-Oise reference might be traceable to a museum, but I can't find anything.
I can't answer any if the questions asked but can throw a couple of examples into the mix. The first:- Silver Tetradrachm Obv:– Laureate head of Apollo right with short hair right. Rev:– PATRAOU, armed warrior on horse prancing right, spearing fallen enemy who holds spear and shield, EM monogram in left field Minted in Kindom of Paeonia, Patraos from . B.C. 335 - 315. Ref:– SNG Oxford 3359 (different dies) 12.646g, 27.5mm, 90 degrees The obverse here is quite a long way off centre though the reverse is quite well centred. You can see some evidence of the concave nature of the reverse. There is some form of mark on the flan ay 06:00 on the obverse that is aligned to the portrait orientation. Was the blank inserted off-centre, struck centrally with the hammer part, creating a decent reverse strike, whilst being such a long way off centre that some facet of die alignment is now evident in the resulting coin. Another coin that I trot out occasionally is the result of an off-centre and uneven strike. I suspect that the blank was inserted and was off centre. The resulting strikes align very well but has resulted in an uneven and off-centred strike. Obv:– Head of Juno Sospita right, wearing goat skin tied under chin. Behind head, Compass Rev:– Gryphon running right; in ex., L. PAPI.; in field, Drill Minted in Rome from . B.C. 79. Reference(s) – RSC Papia 1. RRC 384/1. RCTV 311. Symbol variety – RRC 204. Babelon -. BMCRR - 3.85g. 20.41 mm. 180 degrees
