A GIS Approach for Ancient Numismatics: Spatial Analysis of Antoniniani in Sicily (3rd Century AD)

Abstract

Geographic Information Systems and the use of thematic maps have become well-established tools in archaeology. However, not all the sectors of archaeology still take advantage of these technologies. One such sector is numismatics, where there are still relatively few works on the implementation of coin spatial databases and the related maps. This can be verified both in academic journals indexed in major scientific databases (such as Scopus, ScienceDirect, and Web of Science) and in broader platforms like Google Scholar. In this paper, in an attempt to begin filling the gap, the methodology and results of the creation of the GIS and the Atlas of Antoniniani in Sicily are presented. The second half of the third century ASD is an interesting period because of the socioeconomic crisis that characterized it. The Atlas serves as a useful tool for providing a fresh new insight into the economy and coin circulation during this time.

1. Introduction

Geographic Information Systems and the Atlas production are consolidated instruments used in recent decades in the archaeological field, because they are great for the cataloguing, the management and the analysis of the data in the sector [1,2,3]. They are able to support each step in the archaeological thinking, research and decision system in the field from the creation of digital catalogues [4], to the analysis and calculation of sensibility archaeological areas [5,6].

However, not every discipline connected to archaeology takes enough advantage from these technologies. This is the case of numismatics. Of course, in the articles and monographs related to the topic, maps are produced and used, but they are not dynamic, digital maps, deriving from spatial databases; instead, they are an “instantaneous” and “static” instrument, used just for visualization. Between maps belonging to this second category, the most diffused are present in the studies concerning the production mints, or they are used with the only scope to list archaeological findings. Consequently, maps are often used as an instrument for the coin visualization, without any type of criticism.

GIS is often used and widely used to produce also static and illustrative maps, however its full potential is in allowing a more analytical and interpretative use of spatial data. In fact GIS, more than improving data visualization gives the opportunity to reuse, share, analyze the data; moreover it facilitates the identification of spatial patterns and relationships between archaeological and environmental elements. In this way, the analytical dimension of GIS helps and supports archaeological reasoning and opens up new opportunities for interdisciplinary research.

There are few papers where the combination of GIS and the analysis of coin distribution is aimed at the intra-site study [7,8,9]. In other studies, GIS is used to analyze monetary discoveries according to stratigraphic sequences [10,11]. Also, there are few studies that use GIS and spatial analysis of coins at territorial scales [12,13].

Moreover, on the internet, there are some interesting interactive digital maps, pointing out that the numismatic world is starting to become aware of the potential of GIS for this discipline. Between these, particularly interesting is the Oxford Roman Economy Project [14], where together with coins and coin hoards, as well many other elements, settlements, and activities of the Roman imperial (between the Republican period and Late Antiquity) economy are georeferenced, shown and quantified in order to find patterns of economic integration, regional specialization and changes in trade networks over time as specified in the works cities in the section of the web site dedicated to the literature produced “Oxford Studies of the on the Roman Economy” (just some examples are [15,16,17]). In this way, the study gives useful information about how the Roman imperial economy worked and changed from the Republican period to Late Antiquity. Another example is the Coinage of the Roman Republic Online project [18,19]; in this case, the portal is simpler compared to the side of the Oxford Project, in fact, it is dedicated only to georeferenced coins, hoard collections and their characteristics. The American Numismatic Society developed the Coin Hoards database [20] (https://coinhoards.org/, accessed on 4 June 2025), is an online resource dedicated to coin hoards (mainly Greek and other coins dated between ca. 650 and 30 BCE). The platform includes mapping tools to visualize findspots and contains information regarding coins’ list, typological descriptions, mints and bibliographical references; this is an online resource dedicated to coin hoards (mainly Greek and other coins dated between ca. 650 and 30 BCE). The platform includes mapping tools to visualize findspots and contains information regarding coins’ list, typological descriptions, mints and bibliographical references. This is a particularly virtuous example since the data are structured and distributed as Linked Open Data. Another interesting project is the FLAME project, which combines a spatial representation of maps with the calculation of finding heatmaps [21] (https://coinage.princeton.edu/). Finally, another webGIS related to coins to be cited is the Sacred Coins project [22,23] on Punic and Neo-Punic coinage.

More commons are instead the already cited “static” maps, even if supported by a rich database, as it happen for the website Monnaies de l’Empire Romain/Roman Imperial Coinage AD 268–276 (https://ric.mom.fr/fr/), where are collected and available for consultation different coin types, with information related to their dating and issue, in particular for the historical period analyzed in this paper.

The use of geoinformatics, in particular through the use of spatial analysis, could also be important to understand the relationships and the economic connectivity between different parts of the Roman Empire. By mapping where coins are found and where they are issued GIS can find patterns of circulation and exchange that show how integrated or isolated provinces are, can reflect economic and military choices or can highlight commercial flows.

The aim of this paper, specifically, is to show the spatial distribution and diffusion of coins of the second half of the third century in Sicily. Moreover, the study aims to demonstrate the potential of GIS to promote a shift from purely descriptive approaches toward more integrated and analytical frameworks.

2. Case Study: The Antoniniani Period

The second half of the third century, before Diocletian’s reign (284–305 A.D.), is considered one of the most troubled periods of the Roman Empire. During that time, the common silver coin was the antoninianus, an alloyed coin introduced by the Roman Emperor Caracalla in 215 A.D., weighing 1/64 of a libra, equivalent to about 5.11 g, originally having a value of two denarii but subject to continuous depreciation of its legal value, weight and metal content. The coin owes its name to the emperor Marcus Aurelius Antoninus, surnamed Caracalla, and it is also defined “radiate” because of the bust of the emperor wearing a radiate crown engraved on its obverse [24,25]. The production of antoniniani was temporarily suspended between 219 and 238 A.D.; from Balbinus and Pupienus’ reign (22 April 238 to 29 July 238 A.D.) they were minted continuously until the reign of Aurelian [25,26]. This Emperor, during the spring of 274 A.D., restored the trimetallic system of gold, silver and bronze giving more importance to his radiate silver coin, the aurelianus or aurelianianus that substituted the antoninianus (its weight equivalent to about 4 g being 1/80 or 1/84 of the Roman libra) [25,27,28,29]. Aurelian’s antoninianus did not result in being very popular as it only circulated in the western part of the Empire until Diocletian’s reform in 294 which introduced a new laureate billion coin, called nummus and a radiate neo-antoniniaus. During the last three decades of the third century, the antoniniani struck before Aurelian’s reform circulated almost until the Tetrarchy (in the first two decades of the fourth century A.D.) together with “radiate imitations”, i.e., regional imitations of official coins, also named “barbarous radiate” (a current detailed bibliography can be found in [25,30]).

3. Materials and Methods

3.1. Data Research and Engineering

A large amount of documentation used in this paper comes from historical publications and bibliographic sources, such as journals like AIIN or Notizie degli Scavi di Antichità, monographs or printed catalogues (the complete references are reported in Appendix A). However, most of these sources lack specified stratigraphic sequences. Additionally, a small number of coins that have not yet been listed or are still unpublished are preserved in Museums or Antiquaria. We are conscious that this paper can be enriched with the report of new stray finds or of other coins in Museum or found in archaeological research.

Moreover, third century antoniniani sometimes lose their role as chronological indicators because either they continued circulating for some decades after their being issued or are separated from other archaeological materials thus lacking any contextual information. However, the study of these coins’ circulation could be important to addressing other matters of historical interest, such as the degree of monetization, the use of previous and following currency or the quantification and chronological limits. Thanks to the awareness of the potentials of archaeological contextualization of monetary discoveries, our approach would like to evaluate some data provided by the antoniniani circulating in Sicily.

For the aims of this paper, a spatial database has been created with all the information about different aspects related to the antoniniani in the following fields (

Table 1. Summary diagram of the Antoniniani spatial database.

Field	Description/Vocabulary
Id	Identification number
x	x coordinates
y	y coordinates
Finding place	City of finding
Province
Toponym
Accuracy	- Finding place accurate (yes) - Finding place not accurate (no)
Finding type	- Survey - Archaeological research/stratigraphic context - Coin hoard - Gift to a museum - Private collection - seizure
Numerosity	Number of coins in that location with the same authority, mint and type
Obverse
Reverse
Mint
Issue
Minting date	Chronology

):

• Two spatial fields containing the geographic location of the coins. Unfortunately, in many cases detailed coordinates, which would be desirable for GIS-applications, are not available. In fact, there are cases where only very vague information is available and the record refers to the cadastral municipality as find spot, or we should refer to the toponym location.
• The identification number.
• The findspot.
• The toponym.
• The actual Sicilian province.
• The accuracy, in order to indicate whether the coordinates identified are accurate or, as indicated above, refer to a wider area.
• The find type distinguished between survey, archaeological research or stratigraphic context, coin hoard, gift to a museum, private collection or seizure).
• The number of coins. This information is especially important in case of treasures with some coins of the same authority, mint and type.

We have also registered information about:

• The portraits on the Obverse;
• The Reverse type (and if possible, reverse legend);
• The mint;
• The issue; in this field for “barbarous radiates” we have utilized the term “irregular mint”;
• The specific chronology (mint date).

While this database provides a foundation for studying the circulation of Antoniniani in Sicily, it is understood that it represents an ongoing process. As new discoveries are made and further research is conducted, the database can be continually updated to refine our understanding of the economic and historical context of coin use in this period. This approach not only contributes to numismatics but also offers valuable insights into broader archaeological and historical inquiries.

3.2. Kernel Density Estimation (KDE)

For the analysis of the mint distribution, a KDE approach was used. This method is used in very different application fields [31,32,33,34,35,36] as well as in archaeology [5,37,38,39,40]. It allows researchers to obtain a local and smoothed estimate of a probability density for a point pattern distribution. It is defined as a moving three-dimensional function that weights events (points) depending on their mutual and their intensity [32] resumed in expression (1).

$${\widehat{\lambda }}_{\tau }\left(s\right)=\sum _{i=1}^n{\displaystyle \frac{1}{{\tau }^2}}k\left({\displaystyle \frac{s-s_i}{\tau }}\right)$$

(1)

In it, k is the kernel, which is the shape of the weighting function; s is the point intensity or a quantity identifying event strength; τ or bandwidth is the most important parameter in KDE and defines the width of the weighting function and consequently also how many events are weighted together. For an in-depth study on KDE, see [41].

In this work, the Epanechnikov kernel was used, in order to limit locally the density estimation, as intensity the number of mint was introduced in the calculation and for the definition of the bandwidth a consolidated method used in the literature is used, the 1st order nearest-neighbor average observed distance between events [41].

4. Results

The analysis of numismatic data about the second half of the third century Sicily in a GIS allowed to visualize and explore different kinds of information with the help of several approaches to management of data and spatial analysis. The generated maps, in addition to showing the spatial distribution of the antoniniani on the island, is useful as analytical tools that aid in the interpretation of patterns and densities that may not be apparent from the raw data alone.

The number of coin finds which have been found in Sicily has a great relevance; the results are 107 sites of finding and a total of 1149 antoniniani found there (Figure 1); in a few cases we have only the report of the antoniniani discovered in archaeological sites but without the total number of specimens or the attribution to an authority. We also add, in the maps, 4472 coins which came from the shipwreck in the bay of Camarina even if this big amount of antonianiani constitutes a special case; it is indeed uncertain where the ship of the “six emperors’ coin hoard” was headed, Sicily or other mediterranean harbors, so his story is different from the other kinds of finding [25].

In addition, the data accuracy map enables us to visualize the distribution of findings, concentrated in eastern and central-southern Sicily, especially near Syracuse and Ragusa, along the natural rivers and waterways. These data can offer possibilities to search relations with other archaeological elements, such as infrastructures (e.g., Roman roads) or settlements, and with the topographic features existing in the area and surroundings in a broader interpretative framework. The elevated concentration of coin finds in particular areas could be accounted for by the importance of the places concerned during the second half of the third century.

The first map (Figure 1) distinguishes between accurate and inaccurate data; in particular, it shows how only some records show more accurate information of the place of finding. GIS data are visualized even if we do not have perfect geographic references because the coordinates sometimes are not detailed or they refer to a somewhat vaster area. In the majority of cases, the inaccurate data refer to coins seized or belonging to private collections from the last century; in a few cases, we also consider inaccurate data owing to the absence of particular references to official archeological investigations, even if the area of coin finding, albeit vast, is known.

A more widespread penetration of radiate coins would seem, at the current state of research, to be found in the hinterland of the current provinces of Syracuse and Ragusa and in some investigated archaeological sites of central Sicily as, in particular, in the fourth century Villa Romana del Casale or in the hinterland of Milena (CL). An interesting coin hoard, of 304 AE coins dated from the emperor Adrian to Constantius II, was found in the site of Sofiana, identified as the statio or mansio Philosophiana; it includes over 80 antoniniani, several of which are issued by irregular mints [42,43].

Most of the radiate coins recovered in Syracuse come from the cemeterial area of Vigna Cassia, from the hypogeum Trigilia and from the archeological surveys in Piazza Adda. Palazzolo Acreide (the greek Akrai, in the Iblei mountains) provides an important amount of antoniniani from stratigraphic units. The radiates from the province of Raguse are, unfortunately, in many cases coins from seizures carried out in its territory. For the north-eastern area of the island we can include several specimens from hoards and archaeological excavations in Naxos.

The comparison of chronological distributions of antoniniani provides a more complete overview; however, one is aware that the period of circulation of the coins does not coincide with the time of their issue. Many specimens recovered in Sicilian territory certainly circulated for several decades until the end of the third century A.D., if not even until the following two centuries, as in the case of some antoniniani hoarded together with later coins (e.g., Siracusa–Ortigia, [44] or Naxos [45]).

Taking into account the years of issue of antoniniani found so far, and the bust of the authorities on the coins’ obverse, we preferred to identify four chronological periods: before 260 A.D. with only just over ten coins; the years of Gallienus’ sole reign, A.D. 260–268; the period of Claudius II’s reign, including his commemorative antoniniani, A.D. 268–270; post-270 A.D., from Quintillus to Tacitus. The following four maps (Figure 2, Figure 3, Figure 4 and Figure 5) let us understand in particular the higher presence, in the island, of antoniniani struck under the sole reign of Gallienus, about two hundreds coins in total, and that of his successor Claudius II Gothicus, with an important diffusion of coins, commemorative of his deification after death. This last group is the most representative of the third century A.D., mainly spread in the area of south-eastern Sicily; we just have to remember again that antoniniani, in Sicily, had to circulate some decades after their minting, probably until the beginning of the following century.

A particular case in Sicily is constituted by the coins minted under the emperors of the Imperium Galliarum (from 260 to 273 A.D.) and the imitations of their types (Figure 6); the few coins with the bust of Postumus (260–268 A.D.) are followed by those of Victorinus (268–271 A.D.) and by the even more numerous ones of the Tetricus I and his son (271–274 A.D.). The area of greatest circulation of the coins of the Gallic Empire coincides with that of the emperors of the Central Empire.

The comparison, by means of a histogram (Figure 7), between the types of the most representative reverses of the Tetrici found on Sicilian territory and issued by the two official mints, Treviri and Colonia (=Cologne), and by irregular mints, provides a more indicative picture of circulation in recent decades of the century. On the island, according to the current state of research and from available data, it would seem that slightly more than half of about 300 antoniniani are imitation coins. This suggests that Sicily may have provided landing points or constituted a transit area for trafficking and vessels from Gaul. A relevant case is the conspicuous number of specimens from the so-called “treasure of the six emperors” found in Camarina (see [30]) not taken in count in the histogram due to his particular specificity. The most representative Reverse’s types of the Gallic usurpers, Pax, Spes and Salus, arrive in far higher percentages than the official issue prototypes (as if to show a precise destination towards this area of the Mediterranean). These types are mostly or almost exclusively imitations of official Tetricus I’s coins struck by the mint of Trier (III–V issues, 272–274 A.D.); it is also noteworthy that the number of barbarous radiates exceeds that of the official coins, as in the case of Pax and Salus.

The second histogram (Figure 8) proposed here, relating exclusively to antoniniani struck by official mints of the Empire, helps to better understand the predilection and contacts of the island with particular areas of the Roman Empire, as well as the commercial circuit to which it must have belonged. The relevant percentage of antoniniani issued by the mint of Rome, with 417 specimens, is almost exclusively attributable to issues in the name of Gallienus (260–268) and Claudius II the Gothic (268–270), together with several specimens with the DIVO CLAUDIO legend (post-270 A.D.). Their presence could be noticed, mainly, in the central-southern and eastern areas of the island (Figure 9).

It is also significant the number of issues from the Gallic mints of Colonia and, mostly, Treviri; somewhat insignificant is, conversely, the presence of coins issued by other mints of the Roman Empire, with the sole exclusion of the mints of Mediolanum and Siscia.

Although in this first experimental phase of data collection regarding the presence of antoniniani in Sicilian territory seems to obtain mainly static maps, a different use of GIS applied to numismatics could be the application of the Kernel Density Estimation method. As seen from the description in Section 3.2, the transformation of data into a continuous surface representing the density of distribution, in this case of coins, makes it possible to identify areas of greater or lesser probability of monetary presence and circulation. In this particular case, the idea was to estimate and compare data on the density and probability of the presence and penetration of radiate coins in the Sicilian territory on the basis of the issuing mints. In this way, it is possible to have an idea, albeit a hypothetical one, of the circulation of the antoninianus with respect to the simple place of discovery. In this specific case, above all, the use of the KDE makes it possible to identify territorial areas in which the coin may have had a greater penetration and circulation.

The distribution and KDE of antoniniani according to the mints of issue is well understood by comparing the three maps (in Figure 10). A negative fact is immediately apparent from the comparison of the last map relative to coins for which the mint of issue is unknown with the first two relative to official antoniniani and their imitations. The number of coins for which the mint cannot be determined is higher, as mentioned above, due to the lack of precise data in the publication sites or the poor state of conservation of the specimens.

Particularly interesting is the comparison between the distribution of coins minted by imperial mints, with that of Rome being most represented, and the distribution of barbarous radiates; the antoniniani and their regional imitations circulated in the same places, especially in southern and central-eastern Sicily. The likelihood of coinage penetration in these areas of the island, at the present time of available data, would seem to have been almost capillary based on the assumed density.

Although the imitations constitute approximately 20% of the coins found in the island, and the official mint coins about 47%, the first arrived in Sicily and were most likely accepted as official ones; several of them were also kept in coin hoards, as in Sofiana hoard or in a coin hoard from an unknown site of Syracuse’s province almost entirely composed of barbarous radiates [46].

5. Discussion and Conclusions

This research could contribute to demonstrating the potential of interdisciplinary collaboration between geographical information science and numismatics. It could be used to provide new inputs for the interpretation of coin finds and to boost greater accuracy. In many cases, in fact, the accuracy and the precision of numismatic datasets in this case is not as high as would be desirable for a GIS to build significant large scale models with a high resolution. Despite these limitations, this study focused on spatial distribution rather than chronological analysis. This decision does not imply that spatial data are inherently more reliable than temporal data—both dimensions present significant challenges. However, while the minting date of a coin is usually identifiable, the moment when it was lost or hidden remains sometimes uncertain. For future improvements, the use of other spatio-temporal analysis would strengthen the model and should be considered.

This work is intended to be a preliminary approach toward a broader application of GIS in the territory of Sicily and almost a sort of spur to map the finds on the island in order to have a clearer idea of the monetary distribution at a particular time for the Roman Empire, a period often considered to be in crisis but which for the island shows a very different picture, that of a period characterized by an important movement of men and goods, by an important commercial liveliness, noticeable mainly from the years of the reign of the Tetrici onwards, as demonstrated by the recoveries of imitation coins found all over the island. Important monetary flows would have affected Sicily, a point of passage or even landing on Mediterranean trade routes [25,47,48].

This study investigates other possibilities of the deployment of GIS methods in numismatics and would highlight the distribution of the coins from their place of minting to the places of finding and especially the reasons for which those coins arrived in the island.

However, the results depend on the quality of the data source; we hope that more importance will be given to the protection of data regarding numismatic findings which, in any case, are subjected to greater risks of being introduced into the black market because of their size and sometimes their value. It is also hoped that areas that currently appear to be without numismatic finds or of which we are currently unaware will provide new data that will complete the picture sketched so far for Sicily in the second half of the third century.