Minting of Coins During Gallienus's Reign

[nogoodnicksleft]

The following coin is one of my recent acquisitions.

https://www.forumancientcoins.com/gallery/displayimage.php?pos=-110840

The coin is a Gallienus Tigress type (more commonly described as a Panther), however it has a reverse legend that is normally associated with the Centaur/Griffin types. There is a reverse die match with another coin on lunalucifera.com which has the following description

Obv: [GALLIEN]VSAVG - Head right with radiate crown.
Rev: APOLLINICONSAVG - Tigress walking left. Hints of the officina mark, probably a B, in exergue.
A fascinating and enlightening error. The reverse legend is mismatched from the animal. The animal type is clear and deliberate, yet seems unlike other tigresses. Notice the very well-defined stripes. Let's imagine a possible scenario - The beta workshop is in need of some extra help. They borrow a worker from the delta workshop, who is used to cutting gryphons into the coin dies. The animals aren't that different in appearance, and perhaps he even started engraving the wing before catching himself! (Look at the enlargement - see the diagonal "scratch" across the upper field?) He is unused to engraving the tigress and exagerates the stripes a bit, and also elongates the neck upwards (like the gryphons). He also forgets to change the legend he is so used to cutting, and the die reads with a dedication to Apollo rather than Liber Pater.
The frequency of the stripes on the "panthers" as well as their deliberateness on this coin shows without doubt that the tigress was an intentional type. I believe the tables can be turned on the "panther" - are all of the Liber Pater coins supposed to show a tigress, and some engravers were simply sloppy and presented a more plain, generic feline?
This error type is unrepresented either in the Cunetio Treasure or Gobl, making it a true rarity.

http://www.lunalucifera.com/zoo/liber.html

However the obverse die is different to the lunalucifera.com coin and this raises a number of interesting questions.

Firstly when you consider that the zoo coins themselves are supposed to represent am invoking of the gods in favour of the emperor, how would a coin with a mismatched legend & symbol fit in with a roman society with a system of religion based on complex rituals?. Would such a coin be looked on as bringing extra favour or perhaps the opposite?

Secondly the fact that coins have different obverses got me wondering on how such a thing could happen.

One reason that cropped up when researching the subject of minting was that original observe die broke and was replaced with another. This seems to be the accepted view of how coins with a single die match were struck i.e. one coin at a time until one of the dies broke and was replaced.

Another reason considered was that one or both coins were fakes, having examined the coin closely it seems genuine to me. It and the lunalucifera.com were both sold this year through reputable sources so I’m confident we can discount this scenario.
 
A third possible hypothesis occurred to me, this being that it could also indicate that the striking of coins was done in more assembly line fashion. The idea being that more than 1 observe die was used with a reserve die. So that the first obverse die would be struck whilst the second observe die was being loaded with a blank. In theory this would almost certainly increase the rate of coin production, assuming that the blank on the second die doesn’t fall off after the initial strike.

The fact that these two coins exists, to me indicates that quality control wasn’t such a big consideration during this period. We know from the volume of coins for sale today that many coins were minted during Gallienus reign, so it is logical to assume that coin production rate would almost certainly have been a factor in the mint operation.  Quality control often goes out the window at the expense of speed!.

Any thoughts?

[curtislclay]

The same rev. die being coupled with two or more obv. dies is not at all unusual. There is no need to assume that a particular rev. die was tied to a particular obv. die until one of them became unusable and was replaced. Rather several obv. dies and several rev. dies were in use simultaneously, and the die combinations could change, perhaps when the rev. dies were collected in the evening and then handed out again the next morning.

One detail we do know about the method of combining dies is that two rev. dies were often struck alternately and at rapid speed at the same obv. die. When a finished coin wasn't removed quickly enough, it was often overstruck on the rev. by the second rev. die. A lot of those coins with perfect obverses, but overstruck reverses, survive, which is what led Colin Kraay to his brilliant insight that two rev. dies must have been in simultaneous, alternate use!

When you show a very rare or unpublished coin, it is useful to name the provenance if known. I note that your piece comes from Berk Buy or Bid 191, 29 July 2014, lot 270, and the lunalucifera coin from Gemini XI, 12 Jan. 2014, lot 496. My text for the Gemini coin provides the interesting information that both of these coins came from one and the same extensive private collection of Animal coins of Gallienus. For your specimen a further provenance is also given:  the collector acquired it from Münzkabinett Funk in Bavaria, before 2002 because the ticket is still priced in German marks, not euros.

[nogoodnicksleft]

The same rev. die being coupled with two or more obv. dies is not at all unusual. There is no need to assume that a particular rev. die was tied to a particular obv. die until one of them became unusable and was replaced. Rather several obv. dies and several rev. dies were in use simultaneously, and the die combinations could change, perhaps when the rev. dies were collected in the evening and then handed out again the next morning.

Thanks that is something that I didn't consider.

One detail we do know about the method of combining dies is that two rev. dies were often struck alternately and at rapid speed at the same obv. die. When a finished coin wasn't removed quickly enough, it was often overstruck on the rev. by the second rev. die. A lot of those coins with perfect obverses, but overstruck reverses, survive, which is what led Colin Kraay to his brilliant insight that two rev. dies must have been in simultaneous, alternate use!

It is difficult to image how using two alternative reverse dies with the same obverse die would actually work in terms of speed of production, where would be the best place to read further about Mr Kraays insight ?

When you show a very rare or unpublished coin, it is useful to name the provenance if known. I note that your piece comes from Berk Buy or Bid 191, 29 July 2014, lot 270, and the lunalucifera coin from Gemini XI, 12 Jan. 2014, lot 496. My text for the Gemini coin provides the interesting information that both of these coins came from one and the same extensive private collection of Animal coins of Gallienus. For your specimen a further provenance is also given:  the collector acquired it from Münzhandlung Funk in Bavaria, before 2002 because the ticket is still priced in German marks, not euros.

Ok noted about the provenance I will update the record. Regarding being unpublished I only recently purchased a copy of Gobl, and am still getting familiar with it. Unfortunately my German is non existence so its going to mean a lot of typing in google translate. Perhaps you would be able to clear something up for me Table 20 is where most of the zoo coins action is and I've been looking for a key that explains what the column heading means i.e B1, B2 etc, I'm assuming that it is bust types but couldn't find a key anywhere?

[jmuona]

I believe Kraay's theory has only been published by Curtis, both here on the Forum and in a review he did a short while ago.
However, it is clear that this is what happened. Once a coin is struck, it had to be removed and when this was done another flan was put in and another strike went ahead. It appears obvious this is faster than using only one reverse die. Besides, it also leaves some time for the reverse die to cool. Practically all studies this far show that reverse die consumption exceeded obverse die consumption, so the "rest" may be an additional benefit. A short time ago I showed here an Alexandrian tetradrakhm of Otho which was clearly double struck because the flan remained in the anvil during two strikes with different dies. The odds of finding such a coin if it would require a situation, where the reverse die is changed because of wear are infitesimal. When two dies are used at the same time, such cases will be much more common, as great numbers of coins were struck.
If these matters interest you more, you should get acquinted with Warren Esty's papers on die-count¨estimates and the complexities of die-chains. The ANS library is a good place to get them online.
s.
Jyrki Muona

[nogoodnicksleft]

One other question regarding provenance, should I link the previous ownership of my coin to lunalucifera.com owner as the listing in HJB for my particular lot doesn't actually state this?. Also I never received any copy of the previous ticket from Münzkabinett Funk should I be requesting a copy for my records?

[nogoodnicksleft]

I believe Kraay's theory has only been published by Curtis, both here on the Forum and in a review he did a short while ago.
However, it is clear that this is what happened. Once a coin is struck, it had to be removed and when this was done another flan was put in and another strike went ahead. It appears obvious this is faster than using only one reverse die. Besides, it also leaves some time for the reverse die to cool. Practically all studies this far show that reverse die consumption exceeded obverse die consumption, so the "rest" may be an additional benefit. A short time ago I showed here an Alexandrian tetradrakhm of Otho which was clearly double struck because the flan remained in the anvil during two strikes with different dies. The odds of finding such a coin if it would require a situation, where the reverse die is changed because of wear are infitesimal. When two dies are used at the same time, such cases will be much more common, as great numbers of coins were struck.
If these matters interest you more, you should get acquinted with Warren Esty's papers on die-count¨estimates and the complexities of die-chains. The ANS library is a good place to get them online.
s.
Jyrki Muona

Thanks I'll look up Warren Esty papers.

I don't see how using two reverse dies to one obverse die is going to be any quicker, than using one reverse die to one obverse die (assuming that it is obverse die that is fixed). The actual time taken over the striking is likely to be marginally slower because the first die needs to move out of the way to make room for the second die. Rather it seems to me that it is the reloading of the blank flan is the key action that will dictate the speed of coin production.

I'll be interested in reading about the die studies showing about the reverse die consumption being greater than the obverse. I can't think how heat would play a part, rather I suspect it would be more down to the differences in shape (busts generally being more spherical in shape and likely to be stronger) or perhaps angle of strike could account of it, the fixed dies would always be more or less the same where as the movable reverse die would have more lateral stress by not being perfectly aligned. Or simply the action of the hammer hitting the top of the reverse die, might be enough to weaken the reverse dies so they wear out quicker. The obverse die is farther away from the hammer impact and also has the softer blank flan in between which will dissipate some of the force from the hammer blow.

[jmuona]

Quote:
"The actual time taken over the striking is likely to be marginally slower because the first die needs to move out of the way to make room for the second die."
There would have been two crews working, both with their own reverse die ready in hand. When the first crew took the coin out to whatever heap the second crew already inserted a new flan and went ahead striking.
Jyrki Muona

[curtislclay]

Perhaps you would be able to clear something up for me. Table 20 is where most of the zoo coins action is and I've been looking for a key that explains what the column heading means i.e B1, B2 etc, I'm assuming that it is bust types but couldn't find a key anywhere?

Apparently B means "head radiate right", but Göbl left it out of his Key on pp. 11-12.

That B can then be modified by additions on Göbl's system.

So B1 means "hd. rad. r., one wreath tie hanging free, the other overlapping emperor's neck"

B2 = "hd. rad. r., both wreath ties hanging free"

lB = "hd. rad. left"

B1P1 = "Bust rad. r., one wreath tie hanging free, fold of cloak on front shoulder"

B1P2 =  "Bust rad. r., one wreath tie hanging free, fold of cloak on front shoulder and also behind neck"

Complicated, but once you know what elements Göbl is paying attention to, like the wreath tie difference, you won't have to consult his key, but can simply deduce from his illustrations what bust type he means, since his plates illustrate virtually every variety.

Given the abundant illustrations, you don't need much German to understand Göbl's catalogue. I don't think you would progress far attempting to understand his main text via google translate, since his original German is very wordy and complicated! Computer translating would probably turn it into complete nonsense.

As to the Münzkabinett Funk ticket, it was meant to be included with the coin, so apparently was erroneously discarded. We regularly throw away old tags and tickets that don't convey any useful information. Unfortunately that makes it easy for someone who hasn't done the cataloguing, and hasn't carefully read our own ticket, to throw away interesting old tickets too, assuming they are worthless!

[SC]

"The actual time taken over the striking is likely to be marginally slower"

However, there are still several  reasons that this might have been a viable technique.

First, while the individual strike might take a bit longer, by giving not just the die but the malleator a slight rest means that you would likely increase the overall production, certainly of quality hammer strikes. Thus longer for strikes but higher overall production due to fewer rest breaks and mis-strikes.

Second, by extending the life of the dies you would save on highly skilled labour at the expense of less skilled labour.   We know that relatively few engravers were working at most mints most of the time. Perhaps this technique signifies that their work was often the bottle-neck if not managed properly.

Shawn

[nogoodnicksleft]

Is there any accepted theory on how hot the blank flans would have been before striking ? Would they have been cool enough for somebody to move them by bare hand, or would some type of gloves or tool been used to hold them.

"First, while the individual strike might take a bit longer, by giving not just the die but the malleator a slight rest means that you would likely increase the overall production, certainly of quality hammer strikes. Thus longer for strikes but higher overall production due to fewer rest breaks and mis-strikes.

What about the Observe dies? the hot flan would likely spend longer in contact with it, surely that would make them more susceptible to any stress from heat?

[SC]

The metallurgical studies that point to coins being struck hot examine the crystal structure of the metal.  To affect this crystal structure they need to be very very hot.  Don't touch the coin ouch my gloves are on fire kind of hot.

In terms of stress on the dies I would think that the slightly longer exposure to heat that the obverse die would experience would still impart less stress than the reverse die would experience being the part struck and thus bearing the greatest part of the energy.

Also, having the obverse die in contact with the blank for the longest assumes that the blank was laid on the obverse die and then the reverse die was put on and struck instead of other possible arrangements such as the use of a jig.  For example a hinged jig could in theory even have had the blank in contact with the reverse die longer.  Though of course use of alternating dies makes the use of a hinged jig much less likely.  the point is we can't be 100% certain of the striking arrangement.   

Shawn

[curtislclay]

Kraay's idea is conclusive in my opinion. There is no other way to explain the frequent occurrence of coins with perfect obverses (no overstriking), but reverses of one contemporaneous die struck over another.

[glebe]

Late Byzantine types are often scyphate, and since the curved dies often don't match perfectly they are normally double struck, one strike for each side (left and right) of the coin. It is not uncommon for two different reverse dies to appear on the one coin.
Presumably the reverse die holder dropped the die after the first strike and picked a new (cooler) one.
Roman coins of course were single struck but if you dropped a die you'd pick up a cooler one too.

Ross G.

[curtislclay]

But having dropped the rev. die after striking the coin, why would I pick up a new die and then strike the same coin a second time?

Surely I would have ample time to see that that coin was finished, so had to be removed and replaced by a new planchet?

Hundreds of workers at Rome and in the provinces, throughout the 2nd-4th cent. AD, all made the same silly mistake of dropping a rev. die, picking up a new one, and then striking the completed coin a second time with the new die?

[glebe]

But having dropped the rev. die after striking the coin, why would I pick up a new die and then strike the same coin a second time?

Surely I would have ample time to see that that coin was finished, so had to be removed and replaced by a new planchet?

Hundreds of workers at Rome and in the provinces, throughout the 2nd-4th cent. AD, all made the same silly mistake of dropping a rev. die, picking up a new one, and then striking the completed coin a second time with the new die?

Good point - I was really only thinking of the different obverse/reverse die combinations.

But I don't know that I'm totally convinced by Kraay's idea on the overstruck coins either, which is - what - that a second dieholder whips in and presents a second die before the planchet guy has had a chance to remove the struck coin?

Could have happened I suppose, but it sounds pretty chaotic to me, with at least four people crowded around the anvil. And phenomenally dangerous, even given the lax safety standards of Roman times. I mean, even with only one dieholder it would very dangerous for the planchet guy - with two dieholders it would be far worse - the hammer will keep crashing down as he tries to fish out the coins, and even if he is using tongs of some sort his hands won't last very long.
 
Ross G.

[glebe]

But having dropped the rev. die after striking the coin, why would I pick up a new die and then strike the same coin a second time?

Surely I would have ample time to see that that coin was finished, so had to be removed and replaced by a new planchet?

Hundreds of workers at Rome and in the provinces, throughout the 2nd-4th cent. AD, all made the same silly mistake of dropping a rev. die, picking up a new one, and then striking the completed coin a second time with the new die?

Actually, having thought about it, I suspect that that may have been exactly what happened.

There is a false strike, the die holder drops his die, the coin isn't (in some cases) properly struck, so they strike it again (with a new die).

This of course is just an alternative hypothesis - I think we need to see some of these overstruck coins.

Ross G.

[curtislclay]

I think we need to see some of these overstruck coins.

I showed one, a SAECVLO FRVGIFERO dupondius of Septimius Severus, about a week ago in a thread on this same Roman Coins board.

I must have 30-50 such overstrikes in my own new collection, all purchased since about 1990.

[jmuona]

Here is my Alexandrian tetradrakhm with double-struck reverse.
Beginning of   ( ) on the left side, end of ( )    on the right, both figures visible. A bit difficult to see, but the Kratesis side has been struck in a slight angle.
As to the dropping idea, how could the die axis remain correct if the coin dropped unintentionally back on the obverse die and was restruck with a new reverse die? Certainly they both should be off and a consistently nearly correct position - like in my coin here - suggests the coin remained on the obverse die and got restruck without being "dropped" in between. Stuck to the obverse die, one might suggest. Then we would be facing the much less likely scenario that this happened just at the same time as when the strikers had decided to change dies.
s.
Jyrki Muona

[glebe]

The die is dropped not the coin. But yes, your scenario works - if they did indeed alternate reverse dies.

Ross G.

[jmuona]

Sorry, missed that!
However, dropping the die sounds even wilder to me as it could easily be damaged and the number of dies used is so low that they must have been regarded as valuable items, much more so than a single flan.
Interestingly, I have not seen a single Roman Otho with this error, although a few with clashed dies exist. Not a Galba or a Vitellius for that matter either and Curtis mentioned earlier that the Otho here depicted is the earliest one he could remember. Perhaps we should start collecting information on these?
s.
Jyrki Muona

[nogoodnicksleft]

It seems to me unlikely that the roman minters would have made a large number double reverse strikes in error when using a minting system based of one obverse to two reverse dies. The loading (& unloading) of the hot blank flan will take time and with the two reserve die holders watching on, they are unlikely to missing the fact that the flan loader has not done his part of the job (unless they were just prating around, or hung over from the night before). In general are the double strikes axis 180 degrees to each other (possibly indicating that the die strikers are opposite to each other) 

I suppose if the flan got stuck in the die on the initial striking (too hot & soft) then perhaps a second strike might have been necessary to free it from the die. That could account for reverse double strikes with perfect obverses. Maybe even on striking a flan of an optimum temperature and in the right manner the coin was suppose to spring off the obverse die, making the loading of the next flan quicker.

Another possible solution might be that the two reverse dies are striking several obverse dies but just moving between each obverse die station. This would allow time for flan loaders to remove the coin and replace it with a new blank flan. The die strikers are otherwise engaged so would not be keeping tabs on the reloading. Various configurations sprint to mind but a circular or semi circular might be best as it would allow a centralise heating source to keep the flans hot.

While we are taking about double strikes this Centaur is interesting, while examining the obverse more closer (as a result of this discussion) I just realised there is also an impression of the Centaur on the obverse
 

[curtislclay]

For Kraay's idea to work, one has to assume the flans were struck at room temperature, no pre-heating, and I think practical considerations and modern experiments reported on Moneta-L (Robert Kokotailo) and maybe Forvm too have verified that this is likely.

One also has to assume a rapid pace of striking, maybe five seconds or less between strikes with the alternating rev. dies. It would be interesting to try to reconstruct the system experimentally. What speed could be attained with practice? Would a third man have been necessary to handle placing the blanks and removing the finished coins?

The overstruck die is generally on either the same axis as the original die, or the opposite axis, 180 degrees different.

I see some confirmation for Kraay's idea in the statistics of the Eauze hoard, which contained no fewer than 2451 antoniniani of Valerian I struck from the same obv. die in combinations with six different ORIENS AVGG rev. dies. Those six rev. dies fall into three obvious pairs:

R71, 1184 spec., die axis always 6h
R20, 1045 spec., axis always 12h

R85, 111 spec., axis always 6h
R21, 77 spec., axis always 12h

R72, 3 spec., axis always 7h
R22, 1 spec., axis 12h.

Such an accumulation of coins struck from the same obv. die cannot have been fortuitous: obviously these coins came more or less directly from the mint. Aren't these pairs of rev. dies the ones that were in use simultaneously, meaning alternately, therefore each producing about the same number of coins, and with the die axes of each pair opposite, because the strikers were holding the rev. dies in the same way, but applying them from different sides of the anvil?

[nogoodnicksleft]

Yes those statistics certainly could be a good indication of Kraay's theory, did they record if any of the coins in the hoard were also over struck ?

[curtislclay]

Of the 28,003 coins in the Eauze hoard, I don't think a single one was recognized as having an overstruck reverse.

This mistake did not happen often, and the individual examples are all very rare. Cumulatively, however, there are a lot of surviving examples, clearly indicating that this error was repetitive and must have had a technical explanation.

[glebe]

For Kraay's idea to work, one has to assume the flans were struck at room temperature, no pre-heating, and I think practical considerations and modern experiments reported on Moneta-L (Robert Kokotailo) and maybe Forvm too have verified that this is likely.

Agreed

Such an accumulation of coins struck from the same obv. die cannot have been fortuitous: obviously these coins came more or less directly from the mint. Aren't these pairs of rev. dies the ones that were in use simultaneously, meaning alternately, therefore each producing about the same number of coins, and with the die axes of each pair opposite, because the strikers were holding the rev. dies in the same way, but applying them from different sides of the anvil?

It seems a reasonable conclusion.

But Curtis, you seem to be saying that the striker was also the dieholder, i.e, that these coins were struck as in medieval times, i.e, with the striker holding a single-handed hammer in one hand and a long thin reverse die in the other.

I'm not sure that this would have worked with Roman and Greek coins, which had a higher relief than medieval pennies. I have in mind the Aquanova guy (I think it was) hacking away with a short handled two-handed heavy hammer, although admittedly he was making large high relief replica Athenian tetradrachms.

In other words I think that in official mint operations (as opposed to a solitary forger, who held his own die) there may have been four people involved, a striker, two dieholders (using some sort of dieholder - not their fingers) and a planchet guy.

Ross G.