Die Links: A Tool for the Numismatist

Ancient coins were struck from dies individually cut by artists rather than mass produced from an enlarged master as is the practice with modern coins. Despite all efforts to make dies the same as each other, differences great and small allow each individually cut die to be identified as an individual. Since most ancient dies were not used as a hinged pair and the additional stress of the striking process caused reverse dies to fail much more quickly than obverse dies, it is quite common to find the same obverse die used with more than one reverse die. In many instances the replacement reverse die was close to exactly the same as the one it replaced. However, when the mint was producing more than one type of coin at a time the replacement could very well be a completely different type. Two coins that were produced by the same pair of dies are said to be "die duplicates." Coins which were produced by the same obverse die but have a different reverse die (or the same reverse with different obverses) are called "die linked".

Nine Die Linked Obverse Die Pairs

Why Die Link?

For the most part die link studies will interest students of a narrow area of numismatics rather than the general collector whose aim is to acquire a single representative specimen of attractive coins. However, general collectors can benefit greatly by the studies of specialists and can take interest in owning coins die related to coins illustrated in the primary references. Dies vary according the skill of the engraver and the luck of the moment. In many issues there are dies that are so much superior in artistic execution that their coins are worth a considerable premium over similar items from inferior dies. While these distinctions could be classified as matters of opinion, they certainly result in some collectors being willing to pay more for some artistically superior VF coins than they would for a poor style EF.

Die identification recently played a role in the controversy surrounding the "Black Sea Hoard" which are now generally accepted as modern fake diobols of Mesembria and Apollonia Pontica in Thrace. Very early in the investigation of the hoard the observation was made that the hoard contained only a few coins that linked to previously known dies. The linked coins seemed somewhat different in style from the bulk of the hoard so the question arose if the linked coins had been added to the hoard to add an air of authenticity to the fakes. This lack of die linking was not conclusive in proving the hoard to be fakes but simply brought up a question that needed to be investigated. Similarly, die identity can link coins in question to specimens known to be good (or bad). Unfortunately, fakes made by casting will retain the die characteristics of the original coin used to produce the mold. Cast duplicate coins will share faults that were on the original but will each add characteristics of their own. Collectors need to be careful not to oversimplify the use of die identity in the process of authentication. Die identity is only one of several tools to be used in the process. There certainly is no substitute for experience and common sense. Like the coins in the first image, all but one of the remaining examples shown here are coins of Septimius Severus and his wife Julia Domna. My collecting interests have led me to handle many more of their coins than any other ancient issue. The same points could probably be illustrated by many other specialized collections but here you see images that were available to me. Die studies are easier when dealing with rather rare issues since these tend to have used fewer dies. The sheer volume of material that must be handled to find a die match among Tribute Pennies or Athenian tetradrachms is massive. On the other hand some rare issues are known only from a single pair of dies so all known examples will be die duplicates. A few issues have been studied and published with all known dies illustrated but for many specialities the student will be entering new ground.

The three coins above differ greatly in terms of strike centering and wear making them harder to 'read' but all three were struck from the same die. Often the mint practice was to use a die until it failed due to excessive wear or cracking but in this case the die flaw (1 o'clock) did not progress between examples. It simply serves us as a convenient marker by which we can recognize this die. Occasionally it is possible to establish the order of production of two coins through progressive wear or expansion of a crack. Students refer to this as the 'state' of the die. A new die is called 'early state' while the worn or damaged die is 'late state'. A few examples will be found where a die was retired and recalled to use at a later date but for the most part it will be safe to say that coins that are die linked were in production in the same mint at about the same time. This can be valuable in dating issues that lack TRP, IMP or COS numbers. If an undated reverse was used with the same obverse die state as a dated reverse it is fairly safe to date both coins to the same period. If it is possible to see a change in die detail between the coins, it might be possible to assign an order of striking to the coins but it is important to distinguish die wear from coin wear or strike quality. Here the coin on the left is much sharper than the other two but making such a call from just three coins could be deceptive. If we had another handful of these coins, it might be possible to advance or retract this theory.

More certain is using die identity to fill in missing legends. If a small flan or poor striking has eliminated some significant lettering but the die used is known from a better example illustrated in a reference, it is safe to fill in the blanks and assign the coin full identity. The exception to this would be in the rare cases where dies were recut and altered. This was sufficiently rare as to be ignored for most every day uses. Die identity of these three coins confirms that the center coin was also the rare variety with legend ending IICO even though that part of the legend is not visable on the flan. Obviously a student of these coins would prefer having examples that actually show the variation but that is not always possible when dealing with rare varieties so even coins as poor as these can be worth collecting. Fortunately, this die has some particularly distinctive letter forms in L SEP SEV making good comparison points. It is also necessary to check relative spacing of letters and design points to confirm that all three were from the same die.

Die Identification Techniques

All ancient coins are individuals. Even coins struck by the same dies will differ greatly in appearance due to factors of striking, flan preparation, circulation, burial and damage incurred in the cleaning process that nearly every ancient coin experiences. It is important to separate differences that can be explained from other factors from differences that could only be the result of different dies. Frequently die identities are approached from a negative standpoint. In some cases it is easier to prove that two dies are different than that they are the same. First attention is given to legend placement and spacing. Letters that are unusually shaped and non-matching breaks in the legend provide all that is necessary to eliminate most potential die link candidates. Minor details of the portrait must match and bear the same relationship to the same parts of the adjacent legend. Wear and striking differences make it dangerous to rely on precise measurements but a ribbon that turns the wrong way eliminates a match no matter how similar the portraits appear otherwise. Die cracks are often quite helpful but it is important to remember that they can progress with time. The single most powerful tool in recognizing die matches is a familiarity the student develops with the coins being studied. Known dies become familiar 'faces' and require only confirmation by checking the minor points.

Die Use Patterns

Our example is "Boehringer 703 (V345/R481)". While the reverse die shows the large break that eventually destroyed the die, there is one cataloged online that shows an even later die state and much worse damage. Those who care to search will find it.
All ancient mints did not follow the same rules in the way dies were paired and replaced. Most simple was a system using dies paired for life. When one die failed the pair was retired so there would be no die links - only die duplicates. Some issues were struck by a single working team using one die pair until one die failed. This method produced a simple pattern with each die known to be used with very few other dies. This pattern was popular with large Greek silver and includes, for example, the coins of Syracuse in Sicily catalogued in "Die Münzen von Syrakus" by Erich Boehringer (1929). Since obverse dies tend to outlast reverse dies this system produced a pattern, for example, that can be diagrammed:

O1=R1;  O1=R2;  
		  O2=R2; O2=R3; O2=R4
						O4=R4; O4=R5; O4=R6 etc.

However not all issues were made using such a pattern that got the most life out of each die. Some mints used many reverse dies with one obverse. There is room for more study pertaining to the reasons for these different die us patterns. In most cases the die use pattern was consistent at that mint and time making suspect any individual coin that does not fit into the expected norm. My speciality, Eastern mint denarii of Septimius Severus, usually shows many more reverse dies than obverse - even more than might be explained by the fact that obverse dies lasted longer in use. Next we will examine an example from this group.

Die linked pairs are hard enough to find but finding six matching coins is quite unusual. These six all have different reverse types. Note the two victories differ with the right coin showing victory on globe. I have not found an example of a coin using this obverse with a different reverse die using the exactly same type. This obverse is one of the more distinctive and easily recognized dies due to the weak nose on the portrait. I suspect it would be possible to assemble groups of reverses sharing other dies of this series but these would be less easily recognized when sorting through coins. To find these six, I must have examined 10,000 Severan denarii. Being easily recognizable is a definite benefit for finding matches to a specific die. Some dies last longer than others but with the simple system diagrammed above it will be unusual to find a die linked to more than a few others. A more complex issue requiring several teams working side by side could each keep dies separately and exclusively for their own use or all dies could have been secured in common and issued at random at the start of each day. This random system can produce a single die that is known used with a dozen mates. The Emesa mint of Septimius Severus seems to have employed this shared die system. Oddly, it seems to be more common to find die linked obverses with different reverses than it is to find a die duplicate pair or linked reverses with different obverses. If the coins had been produced from pairs that remained together, the percentage of die duplicates (both sides matching) would be much higher. There is also compelling evidence that two (or more?) striking teams might have shared a single obverse die alternating in a manner that would allow the reverses to cool between striking. This system would be quite consistent with the existance of so many reverses found with one obverse.

Some particularly interesting die links are shown on my page discussing Clashed dies.

It is obvious that this student's interest is the Roman emperor Septimius Severus so all examples are from that period. While the examples illustrate some of the points to consider when seeking die links, no essay could point out all the things that could be an aid or a confusion. Experience is the best teacher. Begin by comparing your coins to each other and all the illustrations available in your reference library.

The Test - Try your skills at die identification

If you have looked at the examples above and been able to see the difference between die characteristics and variations resulting from other factors, you are ready to try my 'test'. These coins are neither the most difficult nor the easiest to identify. Small flans and poor centering make the job a bit harder. Die studies are best done from the coins themselves. Second choice is to work with plaster casts or photographs carefully made to be exactly the size of the coins. While I have tried to transfer the various photos included in this image retaining the proper sizes, small scaling errors may have occurred. Different lightings on the photos may also add confusion. Enlarged photographs allow the viewer to see fine details that would be lost but some difficulty may be encountered by persons wishing to compare them with their own coins.

Examine the twelve coins below. Identify any die duplicate coins and die linked obverse or reverse sets. The photo includes at least one die duplicate pair, at least one obverse link and at least one reverse link. Not all coins are necessarily linked to another. Some coins may be involved in more than one link. (Hint!) Answers are at the bottom of this page.

Click for larger image

An earlier version of this essay appeared in the Celator magazine of December, 1995. Unlike the printed page, the Internet allows easy correction or update of this essay. If there is some point you feel should be made more clear, please tell me. The 2011 update adds some additional coins for the benefit of those who had seen the earlier versions but my photo collection does not allow a completely new set of images. Finding good examples for use illustrating this page has taken a lot of searching.

Back to Main page

(c) 1997-2011Doug Smith

Obverses: A=G; B=H; D=E=F; J=K
Reverses: A=G; C=F; D=E; B=H
Therefore coins A&G, B&H and D&E are die duplicates (both sides match).
Dies of coins I&L are not linked to others shown.