Celtic Coinage of Britain

third edition

Click on coin to see hidden information




Evans' Bookplate


About the Catalogue

Catalogue Organization    (Info)



Catalogue of Celtic Coins


The first section of the catalogue lists the earliest forms of coinage: ring money, currency bars and coins struck in Gaul but imported for use in Britain. The coins manufactured in Britain follow, all assigned to tribes and listed accordingly. The tribal lists are arranged with the uninscribed types first, followed by the dynastic issues (those carrying names of leaders).

Several analyses have been used to assign the British-made coins to tribes. Maps of modern findspots identify the most likely tribe for many types. Other coins are typologically linked to these, suggesting a similar tribal origin. The coins were produced with careful attention paid to weight, metal content and die-cutting. There can be little doubt that the mints were under centralized tribal control. This leaves little room for anomalous issues, and indeed almost all the coins may be easily assigned to individual tribal groups.

Listing the small number of enigmatic types in a special section entitled "Tribally Unprovenanced Varieties" was rejected because it provides no useful information. Furthermore, it relieves students of the series of their obligation of identify the issuing authority for every type.

Overall, the decision to attribute every type has not proven disastrous. In the 25 years since Celtic Coinage of Britain was first published, only a handful have required re-attribution. Yet, some attributions are still arbitrary and others remain the subject of controversy. In these cases, the arbitrary nature of the attribution is discussed in the notes and alternative interpretations given where possible.


Dating the Coins


The chronology relies heavily on metrological and metallurgical studies (111). The weight and gold/silver/copper content of the staters changes over time, providing a yardstick for dating. The silver and bronze coins are often linked to the staters typologically, sometimes by privy-marks.

While the weights and metal content studies provide relative dates for the earliest issues; the sequences of names on the dynastic issues and the increasing Romanization of certain tribal coinages provide them for the later types.

Certain benchmarks provide absolute dates to construct a chronology. The earliest imported coins, derived from the staters of Philip the second of Macedon, provide an anchor point because Philip's dates are relatively well-established in the fourth century B.C. The gradual trend to abstraction yields a type-sequence in the Gaulish series and the appearance of the various coins in hoards, especially the Tayac hoard, assigns dates to the earliest imports to Britain (112).

From the introduction of gold staters in Britain, sometime around 125 B.C., to the end of coinage on the continent after the Gallic War, there was a gradual and simultaneous decrease in weight and gold content. This has been used to construct a time line based on the weight of the Gaulish stater and the percentage of gold, which has in turn been used to date the early British staters (153).

Analyses of hoard contents, the introduction of various types in different parts of Britain and typological studies have all been used to refine the dating sequence. Many of the later types carry images adapted from Roman denarii, these in turn providing dates. With the appearance of inscriptions, some of the coins can be related to historical people and events, and the rest of the coinage fitted in between these benchmarks.




No catalogue of Celtic coins is likely to be complete, and no claims of completeness are made for this one. The 1975 edition of Mack's Coinage of Ancient Britain listed fewer than 550 types (subtracting modern forgeries), this catalogue now lists well over 800 types. Claims that well over 1000 types now exist should be greeted with skepticism. Two problems work against completeness.

The first is the problem of modern forgeries. About 1961, a series of forged staters appeared, and these were not condemned until the mid 1980's. The 1964 edition of Mack's Coinage of Ancient Britain listed several of the forgeries as genuine "new types" and the 1975 edition added many more. The forgeries attained a degree of acceptability as a result of inclusion in Mack's catalogue, to the detriment of collectors, museums, archaeologists and the numismatic trade.

Consequently, the 1989 edition of Celtic Coinage of Britain used a conservative approach. It only listed new types that had received proper (often metallurgical) scrutiny, were die varieties or ancient forgeries of known types, or had been found under carefully-controlled circumstances (mostly archaeological excavations).

Although the work of the Haslemere Forger was suppressed to an extent, the forger's workshop continued making fakes of increasingly deceptive character throughout the 1980s, 1990s and 2000s. Most were silver units, but gold types occasionally appeared.

Other kinds of forgeries came on the scene as well. Cast fakes of Durotrigan silver staters appeared in the mid 1980s. These were immediately condemned and published by 1989, but inexpicably, they continued to fool people in the following decade.

In the 1990's, fantasy-types (which have no genuine counterparts for comparison) began to appear instead of fakes of known types. Given the circumstances, the idea that these could become accepted without intense scrutiny is disturbing.

It will take many decades until all the "new types" reported since 1989 are authenticated properly to determine whether they are genuine ancient objects or fakes made to deceive collectors. Consequently, the inclusion of "new types" in this catalogue continues to be done cautiously. Some genuine material will unavoidably be rejected as a temporary measure.

A special section entitled "New Material" has been added to capture objects reported to the Celtic Coin Index since 1989. This new material has not been authenticated, nor attributed by the author of Celtic Coinage of Britain.

The second problem working against completeness is that of coin origins. Evans' landmark work, Ancient British Coins, appeared in 1864 listing many Gaulish coins as British. Over the next hundred years, cataloguers struggled to delete these from the British lists. Inadvertently, several more Gaulish types were added to the later catalogues.

This catalogue continues the process of deletion of Gaulish types. Where Gaulish coins listed in the 1975 edition of Mack's work have been deleted, a note is included in the appropriate part of the catalogue.

New, plausibly Gaulish types, though found in Britain, have not been added. The bulk of these rejections have occurred in the Atrebatic/Regnan/Belgic and Trinovantian/Catuvellaunian portions of the catalogue. For example, a number of coins have been found in Britain carrying animals with long, spindly legs or with two animals back-to-back. Such images are normal for Gaulish (and some Corieltauvian) coins, but not for southern British ones. Where scholars have argued for inclusion, the type is noted in the catalogue, but not assigned a catalogue-number. They would be listed if evidence of British manufacture could be assembled.


Description of Catalogue Entries


Catalogue Numbers


The entries in this catalogue, unlike most coin catalogues are not consecutively-numbered. There are numbering gaps between the successive types to accommodate future discoveries. The Celtic series is unusual because new types are being continually discovered and the numbering scheme needs flexibility. The 1953 edition of Mack's Coinage of Ancient Britain contained 468 consecutively-numbered coins. By the third edition of 1975, a hundred new types had been inserted adding letter suffixes to the catalogue numbers. Because the suffixes already used letters up to "G" and also because several coins had been shifted to new out-of-sequence locations, the numbering scheme was beginning to break down. It has been necessary in this catalogue to replace the numbering system used in Mack's work. The switch to a new numbering scheme is never a welcomed event in numismatics, so hopefully the current one has enough flexibility for future needs. So far, it has survived through three editions of this catalogue – and the conversion to electronic form on the Internet.

The largest gaps in the numbering sequence have been built into the Atrebatic/Regnan/Belgic and Trinovantian/Catuvellaunian portions of the catalogue, especially for the uninscribed silver. Many of the new types reported over the last 25 years are in this area, and it is expected more will be discovered as the result of the use of metal detectors.

The catalogue number has two parts: a MAIN NUMBER of up to four digits, followed by a dash, followed by a MINOR NUMBER of up to two digits (in the form XXXX-XX). The catalogue thus accommodates a million entries, 0000-00 to 9999-99, the format being fixed to facilitate computerization.

The intention is that the MAIN NUMBER represents a major type, and the MINOR NUMBER distinguishes between die varieties. Naturally, there will be disagreement over the definition of a "die variety" and subjectivity is unavoidable. For most purposes, reporting the MAIN NUMBER is sufficient, but for very detailed studies, the MINOR NUMBER may be necessary as well.

Ancient forgeries are listed along with their genuine counterparts in the catalogue. They are genuine ancient objects, but counterfeits of genuine types. Ancient forgeries are given the same MAIN NUMBER as the genuine type but are distinguished by different MINOR NUMBERS.

Modern forgeries and fantasies are discussed in the Appendix, and catalogued in a separate set of forgery plates. This avoids the confusion that would occur if they were to be listed along side genuine coins, especially if they were not identified as fakes.

In the forgery plates, the modern forgeries are given the same number as the genuine coin, but suffixed with an "F", in the form XXXX-XXF. The existence of modern forgeries is noted in the catalogue entries where appropriate (and now given a link to the appropriate forgery plate). Fantasy coins – those modern forgeries for which no ancient coin exists – are listed in the same Appendix as FANTASY 1, FANTASY 2, etc, and are included in the forgery plates.

New types, catalogued in the "New Material" section, are given MAIN NUMBERS starting with 6000.




Generally, the existing name of the coin, the one used in the numismatic trade, is given. Most of these names are derived from famous hoards, e.g.: the CHUTE TYPE. Others refer to some typological detail, e.g.: the TRIPLE-TAILED HORSE TYPE. A chronological chart of these trade names appears in the Appendix.




Date ranges for types refer to the estimated date of manufacture not the dates over which the coins were in use. The period of use often extends beyond the latest date given. Although firm dates are given, chronology is a controversial topic and uncertainties of plus or minus five years are usually intended. Since 1989, new research has called for some revision of the dates, especially the very earliest British issues. Generally, the revisions have been minor, in the range of 5 or10 year adjustments.




Rarity estimates have been assigned based on a computer study of known examples of the coins. The study included the contents of museum collections, published gazetteers, hoards, major private collections and sales via fixed price lists, auctions and bourse activity. Careful attention was paid to eliminate modern counterfeits and double-counting of individual pieces.

During the last 25 years, many large hoards were discovered, and it was difficult to determine the rarity of certain types. Statistical studies were conducted (113) to predict the numbers of coins found but obviously this approach lacked precision.

In the 1980s, a survey was conducted amongst members of the numismatic trade to determine whether the statistically calculated rarities were accurate. In a few cases, the survey indicated the coins were commoner than the calculated rarity values – in these cases, a lower rarity was used for the catalogue listing, and a comment was added to the notes.

The Atrebatic/Regnan/Belgic coinage was affected by this problem the most, the Durotrigan and Trinovantian/Catuvellaunian series suffered to less of an extent.

The process of refinement has been continued in the third edition of the catalogue. Entries into the Celtic Coin Index and Portable Antiquities Scheme records since 1989 have be used to adjust the rarity estimates. There are many cases in which a lower or higher rarity estimate was warranted. The Corieltauvian coinage received considerable adjustment as a result of the Hallaton Hoards. The rarity estimate of every type in the catalogue was reappraised in 2017.

Although uncounted coins exist and new coins are found daily, the relative rarities are likely to be correct. The uniform rarity scale has been constructed as follows:


 Rarity designation   Recorded examples 
 Common   over 100 
 Scarce   51 to 100 
 Rare   31 to 50 
 Very Rare   16 to 30 
 Extremely Rare   1 to 15 





The traditional AV, AR and AE have been replaced with the simpler designations gold, silver and bronze. All Celtic gold coins have three-component alloys of gold, silver and copper. Some coins are described as White Gold, these have a high proportion of silver (the use of the misleading term "electrum" has been avoided). Many so-called silver coins similarly have two-component alloys, with silver the major constituent. Billon coins are silver-copper alloys, with a high proportion of copper. Copper alloy coins are all listed as "bronze", thought readers are cautioned that many Celtic coins are made of rather pure copper. The denominations listed are the gold stater and its quarter, the silver unit, the silver stater, the silver fraction, the silver minim and the bronze unit (154).




Weights are given in grammes, using two conventions. Most coins are listed with the "typical weight" – this is the weight most existing examples exhibit. Some coins arc listed with the "standard weight" – this is the weight the coin was supposed to have when it left the mint. Standard weights have been determined statistically and represent the heaviest weight for the type, when enough heavy examples exist to determine this (155). If the catalogue listing does not specify which weight is reported, the typical weight is intended.




Diameters are reported in millimetres, and these are typical values. Celtic coins vary greatly in diameter from coin to coin and the flans are often irregular in shape. The largest possible diameter is intended. There has been no attempt to determine standard values, if any existed.




The descriptions of the coins include conventions for naming the objects in the fields. A list of these objects, with illustrations appears in the Appendix.




The designations intended for use by archaeologists and other workers are given before the notes. Those devised by Derek Allen (114) for the imported coins: e.g.: Gallo-Belgic A, have been retained, however those for the British-made coins are out-of-date and have been replaced. A chronological chart of these classifications appears in the Appendix.


Museum Holdings


In the notes, an estimate is given of the proportion of coins held in museums. This entry relates to the Rarity estimate. SOME means about one-third of the existing quantity is held by museums, MANY about half and MOST about two-thirds or more. No entry means the number held in museums is negligible compared to the total known.


Next section – Acknowledgements









Sir John Evans' bookplate





Van Arsdell 1992e, 1993c, 1994a

Sills 2003

(See Van Arsdell 1992e)

(See Van Arsdell 1993c)





Kellner 1970

Sills 2003





Sills 2003





Van Arsdell 1989h, 1994d for example

(See Van Arsdell 1994d)





Cowell 1992

Northover 1992





Grierson 1963

Kosambi 1966





Allen 1970





This section provides information about using the catalogue.


Copyright R. D. Van Arsdell 2017