I agreed to give this talk to try to see the forest for the trees, to speak historically about my more technical numismatic work. The main themes are Rome’s relatively short-lived cast coinage a phenomenon starting at or just before the First Punic war and ending less than 60 years later during the Second Punic, or Hannibalic, War, why it was created and how it was used, with a particular emphasis on religious sanctuaries. To do this I’ve been very selective with what I’m sharing to try to give a big picture using broad brush strokes, but as I truly love the trees as much as the forest, please feel free to ask follow ups in the Q&A.

If you would like a copy of my slides to follow along now or for future reference the QR code will direct you to the appropriate post on my blog post, password WashU. I’ll remove the password after this talk is delivered and at the same time post a version of my speaking notes. They won’t be a transcript but rather a record of what I hoped I might say on this event. If you’d like to enter the website address directly just go to livyarrow.org and it should be the first post.

While those of you who wish to do so connect, I share an image of a terracotta votive found at Nemi now in the Villa Guilia. It is believed to be a depiction of the temple of Diana there as it may have existed in the 2nd Century BCE. At very least it is a lovely aid to our imaginations for the now lost setting for many of the objects well see today.

As a historian I cannot over emphasize the importance of provenance for coin finds, not just authentication but also evidence for historic usage through deposition and proximity to other finds, in essence to treat coinage as the material cultural and archaeological evidence it is. Analogy about epigraphy, in the past too often studied from catalogues not from the original findspot and culture of display. There has been much correction of this in epigraphy through a wider spatial turn in classical scholarship over the last 30 plus years.

The materials of Nemi were excavated by the British Ambassador to Italy in the mid 1880s and his portion of the finds are now on deposit in Nottingham UK thus accessible for study along with some documentation of their find context. They are thus an ideal case study to try to help me answer some of my bigger questions around the uses of money in the mid 3rd BCE Italy. I began collecting metallurgical data from aes grave in university museum collections here in the US in 2023 to justify a proposal to access and study the Nemi assemblage in Nottingham.

Through a strange serendipity this year was also when Italian sanctuaries as places of pilgrimage hit the news with splashy images of the discoveries in the hot spring pools at San Casciano dei Bagni. When the bronze statuary and other votives began to come to light the incredible artistry and volume of artifacts inspired a few skeptics to accuse the archaeologists of planting the objects, as they seemed to completely up end our ideas Hellenistic Italy, at least to those unaware of how much has been lost through looting and just sequential reuse and refashioning of sites over time.

While I would not stretch too far the parallel between San Casciano dei Bagni and Nemi they were active in use with similar populations in the middle and late republican periods. And indeed, there are a number of other somewhat similar sanctuaries which I’ll discuss at the end of this talk.

While I have lots of big questions some of my biggest around the emergence of Roman coinage center on two main themes

read slide

The definition of money is slippery and disputed. Hollander in 2007 wrote a wonderful book on the nature of money in the Later Roman Republic to try to bring some of the modern scholarly perspectives to bear on the Roman economy. He uses a common definition of money and shows the diversity of money in use by Romans in the first century BCE. 

Read slide, Discuss.

I would suggest that we consider monetary system a form of intangible cultural heritage.  At times we have physical manifestations of money or artifacts relate to the systems, but the socio-cultural evolution of money as concept need not be physical.

Prior to Hollander and even today quite often scholars implicitly approach the study of ancient economies and specifically monetary objects like coins from a teleological perspective, one in which we presume progress from a state of primitivism to complex societies or “civilizations”. In essence, I’m asking you today to interrogate with great skepticism three common underlying assumptions.

Read slide.

In essence, I’m suggesting that even if we do not understand the nature of an ancient economy or a scarcity of evidence for pervasive use of physical money, we should not assume that a society operated with ”only” barter. Barter itself is a romanticized concept associated with the primitive other as opposed to our own ‘sophisticated’ systems for measuring value, storing wealth, make payments and account for all these and more transactions. Thus my proposition that money can be a form intangible cultural heritage.

This type of primitive to sophisticated intuitive logic for discussing material cultures is pervasive far beyond money and coinage, and it may be easier to conceptualize and recognize the logical fallacy when looking at art rather than economic instruments. If I asked you to date the figure on the left as compared to the figures in bronze in the photograph, you would probably use stylistic seriation to hypothesize that the figure on the left if far older. It has far less detail for the body or facial features, few attributes and no attempt at detailing clothing, let alone conveying the potential for complex dynamic movement in the human form. When these objects were excavated at Nemi by Lord Saville, he guessed that the figure on the right might be Egyptian, and when I got to hold these objects when studying the Nemi materials in Nottingham, I too assumed that the figure on the left was evidence of older cult activity at the site.

And yet, metallurgy and the limited findspot reports suggest this assumption is very wrong indeed.

Based on surface pXRF readings, the figure on the left is heavily leaded bronze, similar to the figures on right. Bronze is a mixture of Copper and Tin and in the height of the bronze age in the Mediterranean world there were very precise recipes for ‘good bronze’ maximizing its strength for tools and weapons. Typically, this might have been about 10% tin to 90% copper, but of course I’m speaking in broad brush strokes.  Bronze is a true alloy but not a single formula by any means.

Even in the Bronze Age Mediterranean peoples shifted their bronze recipes to achieve not only different characteristics of the metal, but also different colors and esthetic effects. In the Iron Age, as the use of bronze shifted, so did the recipes. The third ingredient here is lead and in apparently large amounts. The typical reason to add lead to bronze is to increase in fluidity when pouring molten metal, but this ‘benefit’ is only realized in amounts less than 4%. Over 4% leaded bronze is no longer a stable unified body, but rather an admixture that cools erratically and produces highly inconsistent results.

Lord Savile in neither the Italian journal publication nor his own museum catalogue differentiates the find spots. I believe based on his summary all these objects were found intentionally deposited directly in front of the South front of the Temple, the side facing Lake Nemi. I am hoping to find Lord Saville’s original letters and journals relating to the excavation in the Notthinghamshire archives, but thus far all reports indicate a single bronze votive deposit.

Lead was cheap and versatile materials. Some have compared it to the plastic of the ancient world, seductive in its innumerable applications while silently harming the environment and causing unrealized health problems.

The ready availability of lead in the ancient Mediterranean is directly tied to the silver production. The more silver that is extracted from the mines the more lead ingots can be exported for use in building and plumbing, cosmetics and the arts. Untold purposes. After the Hannibalic War Spanish lead becomes exceptionally common but it was already available in the 3^rd century from mines in Macedonia and Greece and widely traded. Rarely can we trace the exact origin of the lead because of extensive reuse and remelting of the materials.

The high levels of lead in the Nemi votives is there as a filler. It is cheap and heavy and can be added in to stretch whatever collection of outdated and unwanted bronze objects are being melted down to create these small gifts to the gods, thus driving the sanctuary economy while preserving the pax deorum.

Such high lead levels are only practical for small cast objects. If one tried to beat out such metal say for armor or a cooking pan or even a mirror, it would immediately shatter. Such an admixture has little to no flexibility or tensile strength.

We’ll circle back to the question of accuracy of these pXRF numbers, and the short answer is not very, but the overall pattern is convincing and there is no doubt these objects are rightly described as heavily leaded even if we should take specific numbers with a grain of salt.

Let’s now shift our attention back to the origins of Rome’s coinage as part of the existing italic monetary landscape in the 3rd century BCE and how all the Nemi finds together can help us speak with more accuracy about this historic development. We have two more presumptions that have for far too long dominated our discussion of Rome and Italy’s pre-existing monetary cultures.

First, [ read slide]

While I am not all the way to a Peter Wiseman level of skepticism on our literary sources, that sees little knowable before the production of Roman literature in the 3rd century BCE, on which I highly recommend Feeney’s work, I do as a historiographer has a strong sense of the Romans concern to create a plausible coherent narrative of events in deep time. Both Livy and Pliny try to piece together observable physical monuments and their own literary and oral traditions to tell us how Rome ended up the way it did. We can respect their methods but we need to be skeptical of they’re ability to truly know the unknowable or even understand the evidence the do have. I’m happy to discuss the whys and wherefores of my skepticism in the Q&A.

Second, [read slide]

This second one is more at the heart of this talk and will take more unpacking. I also have many sceptics to my view that heavy italic money could be a fiat currency, but I will do my best convince you through out the rest of this talk.

Aes Rude are materials like those you see in the upper righthand corner of the slide. They are unshaped pieces of metal that are presumed to be bronze that do not seem to conform to any particular standard and usually without design.

There are some examples that have rough designs or more intentional shapes like discs or rectilinear ingots, but these are the exception. They are found throughout the Italic peninsula and are believed to have served as money from the 6th century maybe even the 7th century onwards. The Burnett Hypothesis explains the strangeness of Rome’s very first coins, called aes grave, or heavy bronze, as an attempt to integrate the logics of Greek coinages struck in S. Italy with the Indigenous monetary forms. I am largely a supporter of this hypothesis as long as we do not presume that the heaviness relates to intrinsic value.

That large coin at the bottom of the slide weighs a full Roman point about 324 grams. Or about the weight of a can of soda after you’ve had the first gulp, or a smallish cup of gas station coffee to go. You can hold it in one hand but it is going to more than cover your palm. The great confusion about this monetary innovation is that is seems so deeply impractical. Why make it this way?! The standard answer is because it must have a lot of bronze in it to be worth anything.

Our physical evidence of how coins were used together suggests this cannot be correct. On the screen I show you parts of the Ardea hoard recovered in 1940. The coin is comprised of the earliest Roman bronze coins. On the left you see more of the aes grave. On the right we see small struck bronzes made at almost the same time also by Rome. We call this coin a litra but like aes rude and aes grave this is just a modern term we use for our own convenience and bears no relationship to what the Romans would have called them. The struck coins weight about as much as two nickels. Each. It would take 32 or more of these little struck coins to equal one of the largest of the cast coins. And, yet here they are together in the same context being saved or ritually deposited as similar objects of value. We cannot reconstruct an exchange rate but we can see that the two types of coin were made at about the same time and used in similar ways.

If we leave Latium and travel into the Mountains of Samnium we see the same thing. Cast and struck Roman coins together, AND alongside the many numerous bronzes struck by Roman allies in Campania during the course of 1st Punic War to support that war effort. Struck and cast coins are found and used together by people comfortable and familiar with both objects as forms of money.

Much of my work has been inspired by the publications and observations by Jaia and Molinari based on their archaeological and numismatic expertise. They conclusively demonstrated a correlation the first Roman cast coinage between Roman investments in their seaboard defense network post Pyrrhic war., an investment that looks very much like an investment in an anticipated conflict with Western seapowers, namely the Carthaginians. Most recovered aes grave where context is known seems most likely to come from religious contexts, and we see aes rude reported in archaeological contexts as a single piece as part of a grave assemblage. Molinari’s careful review of finds from Lavinium showed aes rude, a Roman currency bar fragment, struck bronze coins of Campania and Rome again dating to 1st Punic war all together under a collapsed roof of a commercial space off the forum. This helps us clearly see that all these monetary objects did function as money and could do so together in Latium in the middle of the 3rd century.

The two sides of this slide taken together help us see clearly that the aes grave phenomenon does not evolve into struck coinage but is a parallel phenomenon and that it appears in our archaeological record on the eve of the 1st Punic War, not earlier.

However, my questioning of the idea that aes grave have intrinsic value based on their metal composition did not in fact grow out of their co existence alongside struck bronzes, but rather pXRF testing of aes rude published by a number of Italian scholars, perhaps most importantly the finds from Ghiaccioforte. This small oppidum North of Rome is believed to have been destroyed the Romans c. 280 BCE and never re inhabited. The site had a vast quantity of aes rude across the settlement in different contexts. The 2007 published results of surface analysis showed little to no homogeny in the elemental composition of these pieces of italic money. If these objects traded based on weight how could everything of the same weight have the same intrinsic value if the individual objects are made up of radically different ‘recipes’. After handling a great deal of aes rude in museum collections and most especially the Nemi assemblage of such materials, I can say that I personally cannot guess the likely metal content of a piece of aes rude based on touch, feel, and looks, and I significantly doubt anyone using these pieces as money would have been able to do so. After looking at this data my only reasonable hypothesis seemed to be that the metal content didn’t matter only the weight of the object. That in fact aes rude was itself primarily a fiat currency, that is a symbolic form of money that derives value from its shared cultural significance not how the material can be re used. The question of what this means for aes grave and the burnett hypothesis has been exercising on my mind now for a number of years.

The logic of worth by weight is baked into the Latin from an early period. Kroll made the linguistic case for the correlation between aes grave and aes rude monetary cultures very clearly in 2008 believing that these objects have intrinisic value.

 My first approach to this problem was to re analyze the metrological data for the earliest aes grave. Previous scholars from Haberlin to Thomsen to even Crawford relied on simple averages to determine weight standards with out statistically considering overall shape of the data and its variability. They also did not look at denominations in relationship one to another.

When Rome starts making cast coinage on the eve of the 1st Punic war a base-12 denomination system appears fully formed like Athena from the Head of Zeus, or Minerva from Jupiter if you prefer.

This denomination system is by far and away the most complex and nuanced known in the Mediterranean at this time. NO other culture seems to have felt the need to make their money so granular in its available units and very few indeed used any sort of mark of denomination like this.

And, even more curious, the small denominations are vastly too heavy compared to the larger denominations. If these coins had intrinsic metallurgical value then the smallest coins should have been hoarded and prized for containing more of the desirable metal. There is also incredible variation in weights. That means little to no concern to regulate the weight of any individual specimen. And no expectation that users will care about such weight discrepencies and instead trade the objects on face value.  In fact fundamental to any coinage denomination system is the expectation that the coins will be accepted at face value. The Romans are communicating by the marks of value how those using their coinage should use them and that is not by putting them in a scale to weigh them.

Obviously my next research question was to wonder if Roman aes grave had the same wildly erratic composition as the aes Rude like we saw from Giacchioforte find. I began with surface analyses and measured museum specimens from Princeton, Yale, and Rutgers. Some 400 readings from well over a 100 individual specimens. BUT such objects are highly at risk of forgery not just now but as far back as the late 18th century! That preliminary work in the museum helped me get to know the likely shape of the data and convince the curators in Nottingham to let me analyse the Nemi material. What you can see immediately on the screen is that while there is variation in individual specimens, notice the spaces between the green dots and the whiskers on the box and whiskers graph, nevertheless the variation is within pretty well-defined margins and with only three dominant elements always in roughly the same ratio. Unlike aes rude, Roman aes grave conforms to a recipe and that recipe is full of lead just like those votives we saw earlier in the talk.

In fact, when we overlay the readings of Nemi figural votives and the aes grave found alongside it, we see they look very much like they are made from the junky admixture of metal.

This to me is hands down the most exciting part of this work and the backbone of my grant proposal due this coming Monday! I believe this connection and the overlap in find spots, namely in religious sanctuaries in the same deposition contexts, maybe our most significant clue to why the Romans chose to create a cast coinage alongside a struck coinage. It may even be that these objects were manufactured together, but for now that is pure speculation.

So now I’m going to get even more technical but try to keep it accessible. When I started my surface analyses using pXRF many, if not most, in the numismatic community rolled their eyes and pointed out all the flaws in the pXRF studies of silver coins and how one cannot know the interior of an object from the surface. Some even refused access to artifacts because they believed I might mislead the public based on my findings. Nevertheless, pXRF machines are regularly using in field archaeology to provide preliminary analyses and my colleagues in bronze age metallurgical studies use pXRF as the backbone of funded multimillion dollar grants. pXRF has limitations and we need to be keenly aware of what causes variations in surface readings and how we interpret this data. Hence my focus on bulk analysis of pXRF results. Nevertheless, the patterns in this data has been the first step in justifying great investment in other forms of much more expensive, time-consuming scientific investigation.

As I mentioned earlier when you add so much lead to a copper-tin alloy it stops being a true alloy and is instead an irregular admixture. No one reading will tell us the recipe used in the original casting. Not only do we have to worry about changes to the surface through environmental effects, especially leaching away of copper from exposure to acids, but the internal structure will have different chemical make ups at different depths dependent on the cooling process after casting. Thus even shallow drilling and wet chemistry cannot provide accurate answers regarding the composition of the original molten metal. Moreover, with uniquely provenance artifacts like those from Nemi, drilling would raise serious ethical questions.

I was excited to meet Adrian Hillier and his colleague Sayani Biswas of the Appleton Rutherford laboratories in the UK and hear about their experimental used of muonic Xrays to look inside cultural heritage objects. Adrian is a key developer of this new technology. You may have heard of neutron activation techniques where a particle accelerator is used to bombard an object with radiation and to reading its total element make up. This older technique leaves the object radioactive for months if not years, but still has its place in scientific investigations of cultural heritage.

Muonic x-rays are also produced from a particle accelerator beam line. We capture muons and target them at the object up stream of the neutron activation stations. We can control the moment of the muon we shoot at the target and thus control the depth of penetration. The muons are captured by the neutrons within the individual atoms of the object and we then can read their decay rate using hyper-cooled germanium detectors. This scientific technique is only available at four laboratories globally and only the UK and Japanese facilities accept proposals from researchers outside their own facilities.

I cannot emphasize enough how much this is new science and very much an experimental technique. The physicists actively developing this technology say they like working with cultural heritage researchers like myself because we ask questions and present materials that are well outside the traditional sciences. I come to the project with a unique challenge and without pre-conceived notions of what is or is not possible. They tell me this lets them be better scientists in making the technology do more sooner. My first grant proved that we could in fact overcome the challenges of the heavily leaded material and through calibration and longer experimental times read the composition under the surface of the coin at different depths. But even just processing and interpreting the results after we achieved an appropriate research protocol took significant time and effort. The code of the software to interpret the readings from the sensors was being refined and further developed in the same control center side by side with us running our first experiments.

These experiments provided proof of concept and while I only show you data from two specimens and two depths each here, this first grant allowed me to successfully apply for a second.

This form of experimentation is time consuming and expensive, but I think it is worth it. On the screen you can see the results of multiple pXRF surface readings of the same Nemi aes grave in orange and compare this to the subsurface MuX readings on the left. The objects still seem very heavily leaded and copper content increases, just as our modeling anticipated. After my six days at MuX this coming May on my second grant, we should have a spread of data from at least three specimens of all Rome aes grave series.

This should let us confirm my hypothesis that there is a relatively stable recipe, but also see if this recipes shifts over time or even perhaps between denominations in series. We do expect small objects to cool differently than larger ones effecting how we can model the interior structure, and thus reconstruct an original recipe.

Our goal is to be able with an original recipe to being experimental archaeology wherein we reconstruct how exactly these objects were cast. We don’t even know what material was used for the molds! Once we can replicate casting we can cut a replica in half and better model the effects of various recipes on the cooling at different depth and determine what might have been the ancient manufacture process.

Additionally, I hope to test the bronze votives in a similar manner to be able to make better direct comparisons. MuX is going offline for about 6 months or more in early 2027 for upgrades which should allow for faster readings with more sensors, and this will empower isotopic analysis and comparison. This is theoretically feasible now but too time consuming with present equipment to be feasible. Isotopic analysis would be most useful for comparison of objects found at Nemi to determine if they may have had the same source metals even if those metals were heavily reused and re-mixed.

So if the aes grave wasn’t just an evolutionary stage on the path from primitive money to fiat currency, and was far more time consuming and resource intensive to manufacture why would the Romans have bothered to invent and maintain it as a form of money particularly during a time of intense warfare in their first conflict with Carthage and overseas campaigning. Some scholars of the third century wondered if they are really money at all but perhaps ‘only’ religious votives or scale weights. We’ve seen earlier that they are found in commercial contexts and are regular enough to be scale weights, but is there any merit in the religious connection. Can we save the baby even as we throw out the bath water?

The money displayed on the left are a small selection of those discovered at the sanctuary of Apollo at Vicarello. The display was created almost 30 years ago in Rome, although the newly appointed curator tells me she wants to take it down. It encodes for the casual viewer that evolution logic so pervasive in numismatic studies and which I’ve argued so strongly against. It does, however, do a good job of giving a sense of massiveness of this find. While workers expanding the bathing area in a catholic hospital in 1852, disrupted a surface in hot springs that had sealed the ancient deposits in the spring. With the release of pressure, literal tons of metal began shooting out of the spring within the hospital complex and it was quickly realized they dated to Roman period. Workers were then tasked with recovery of the materials for church working in water well over 120 degrees with little to no ventilation, an almost unimaginable job.

We see a shift in votive culture from bronze monetary objects at its height in the later 3^rd century to a revival in Augustan age with silver votive vessels.

Votive deposits in water are known from other Italic sites. The lake of the Idols was a source of innumerable finds that have been dispersed on the antiquities market over the sanctuaries in the early 19^th century. I mention it primarily to help you see this religious phenomenon was not particularly unique and that each site has a a slightly different if related culture offerings, even as there is significant continuity across the Italic peninsula in the forms of votives.  The 2003-2005 surveys brought to light a massive amount of aes rude and just a handful of coins.  I’ve not seen any reports yet of what these coins were or what the archaeologist mean by aes signatum in this context.

Finally, I want to circle back to San Casciano Dei Bagni. I am deeply intrigued by the phenomenon of weights being inscribed on bronze statues as presented by Mattia Bischeri at the 2025 SCS. I am watching these excavations and the reports of these inscriptions closely as I believe this may give us an important clue to why aes grave was so heavy even as the metal itself need not have been particularly valuable as metal.  The weight may have been preferred for religious contexts. 

This work will continue and hope in future to share more answers…