return to main index

Re-attribution of the Anglo-Gallic Deniers of Edward I/II: The Denier au Leopard, first type, and the Denier a la Couronne

copyright 1999 by Steve Ford and Marshall Faintich


In 1285, Edward I of England sent the English master of the mint, William de Turnemire, to Bordeaux for the purpose of issuing a new and better quality money for Aquitaine. This new money, the "moneta nova" was delayed until 1291. Unfortunately, poor records, a closing of the mint for eleven years, and the accession to the throne by Edward II, all contributed to confusion about four known types of deniers and two types of oboles that comprise the series of coins addressed in this article.

The "moneta nova" of 1291 came to an abrupt halt in 1294, when the duchy was confiscated by the French, and Bordeaux was occupied. In 1305, two years after Edward I was restored in Aquitaine, the minting operations were resumed. Edward II became king upon the death of Edward I in 1307.

The current standard reference, "The Anglo-Gallic Coins," written in 1984 by E.R. Duncan Elias, revamped many of the coinage attributions compiled by L. M. Hewlett in 1920. However, this particular series of coins presented unexplainable dilemmas for Elias, even though his ordering and attribution of the series utilized additional research and analysis to improve the conclusions of Hewlett.

Hewlett presented doubts that any coinage was struck in Aquitaine during the reign of Edward II, but Elias refers to the research of documentary evidence by Mary Rechenbach that clearly demonstrated that deniers, and perhaps oboles, were struck between 1307 and 1324. Nonetheless, Elias provisionally attributes this entire series of four deniers and two oboles to Edward I, but withholds the possibility that some could have been continued by his son.

The discovery of a fifth type of denier by Steve Ford, combined with a plausible dating scenario based upon research by Marshall Faintich, yields a new, simple, and logical ordering and attribution for the series.

The Coins

The first Denier au Leopard (Elias 18), depicts a leopard above AGL' with an epsilon below, and another epsilon in the first quadrant of the reverse cross. Although Elias cannot explain the epsilon symbol, and he does conclude that it is not part of the legend, as the epsilon is replaced by other symbols, such as a trefoil, on the Denier au Leopard, second type.


The next denier, Elias 19, has the reverse epsilon in the second quadrant, and is extremely rare. Only two examples are known. The third denier, Elias 20, has the epsilon returned to the first quadrant, with a crescent added in the second quadrant. Moreover, the legend below the leopard is changed to AGI. E20 is a fairly common variety.


E21 and E22 are oboles that correspond to denier types E18 and E20, respectively. There is, however, a transitional sub-type, E21a, that reads AGI' instead of AGL'. Finally, there is a related issue, E23, the Denier a la Couronne. This coin has a crown in lieu of the leopard, but like E20, has AGI in the legend with an epsilon below, and like E18, has only an epsilon in the first quadrant of the reverse cross.

Elias attributes E18 and E21 as the "moneta nova" of 1291 that was struck until the closing of the mint in 1294. And he attributes E20 and E22 to the reopening in 1305, and postulates that this type continued into the reign of Edward II. Due to the rarity of E19, Elias speculates that this coin was struck just before the mint was closed in 1294.

E23, also very rare, presented a dilemma for Elias. He attributed it to Edward I, but conceded that it might have been struck well into the reign of Edward II, or perhaps even between E18 and E20, as it has the reverse of E18 and the AGI of E20.

The Fifth Denier (E20bis)

The discovery of a new denier type, E20bis, simplifies the attribution sequence. This coin seems to parallel E19. E18 has an epsilon in the first quarter; E19 in the second; E20 has an epsilon in the first quarter and a crescent in the second; E20 bis has an epsilon in the second quarter and a crescent in the fourth.

E20bis View coin scan

In the case of E19, Elias thought that the movement of the epsilon from the first to the second quarter was intentional and indicated a new and distinct issue. If we follow his reasoning for E19 then E20bis should be similarly interpreted. This interpretation, however, complicates the situation enormously by raising a number of very perplexing questions regarding why and when these changes were made. For example, why would the epsilon be moved from 1st to 2nd quarter; then when the crescent is added, why is the epsilon moved back to 1st quarter; then once again to the 2nd quarter, and finally back again to the 1st quarter when the crescent is removed (E18, E19, E20, E20bis and E23 respectively)? Moreover, given the dating Elias proposes, where can we fit these various issues? If, however, the movement is explained as unintentional, then the situation becomes greatly simplified. Let us now examine this possibility and the conclusions that follow from it.

Upon close inspection, both the Elias specimen of E19 and the Ford specimen of E20 bis look like die rotation errors. E18 has its epsilon facing either upright or diagonally towards the center in the first quarter, as does E20. E20 has its crescent facing left in the second quarter. Curiously, E19 has its epsilon facing diagonally downwards in the second quarter. Moreover, E20 bis has its epsilon facing downwards in the second quarter, and its crescent facing upwards in the fourth quarter. These two seem exact parallels of one another. Finally, look at the legend: there is an initial cross at 12 o’clock and an X (of DVX) at three o’clock. The epsilon should be in the quarter immediately to the left of the cross. On both E19 and E20 bis the epsilon is immediately to the left of the X and facing towards the X (just as it should face towards the cross if correct). If one were to keep the legend stationary and rotate the field 90 degrees counter clockwise, E18 and E20, respectively, would be constructed, including the correct orientation of the epsilon and the crescent. The mistake would have been very easy to make for an illiterate die sinker given the similarity between the cross and the X.

Since both E19 (2 known specimens) and E20 bis (1 known specimen) are exceedingly rare, the coincidence seems too great to go unnoticed. One would have expected to orientation of the epsilon and the crescent to remain the same as in E18 and E20 if the change were intentional. As this is not the case, one may conclude that both coins are due to a die rotation error.

This contention is supported by an examination of the oboles of this series. The common denier E18 and E20 both have corresponding oboles (E21 and E22 respectively). However, there are no known oboles that correspond to either E19 or E20bis. This could simply be due to the exceptional rarity of these two coins, or it could support the contention that these issues are indeed die rotation errors. The key to this interpretation is the obole E21a that has the reverse of E21 but the obverse of E22, providing a direct link between the two. This would seem to suggest that from a perspective of design development there is not a place in the sequence either E19 or E20bis as intentional issues.

With the attribution of E19 and E20bis as die errors, then only three denier issues remain: E18, E20, and E23. And there are three important dates: the "moneta nova" of 1291, the re-opening of the mint in 1305, and the accession of Edward II in 1307. But these events do not fully explain the three coin designs. If Elias were correct in his speculation that E23 could logically fall between E18 and E20 based upon design considerations, then why would the crown replace the leopard when the mint was re-opened, and why would a crescent be added?

Perhaps the answer can be found elsewhere.

"Symbolic Messengers of Medieval Man"

A review of historical documents makes it very clear that medieval man often viewed astronomical events as omens or heavenly messengers. It seems quite plausible that those who minted coins would use symbols that were easily recognized by the largely illiterate populace to remind them of these events and thereby reinforce their own divine right to rule.

In "Symbolic Messengers of Medieval Man," Faintich presents the basic contention that many medieval coins contain such astronomical symbols, and that these symbols often refer to actual celestial events. Thus, the presence of one or more of these symbols associated with a verifiable celestial event can help us to attribute and date many questionable issues. In his analysis, a very strong correlation is shown between the verifiable occurrence of significant astronomical events and the appearance of representative symbols on coins. Thus, although the correlation in any specific issue could be disregarded as mere coincidence, the overwhelming number of apparently meaningful cases, is strong support for the general hypothesis. Among the symbols so analyzed are the epsilon (comet) and crescent (lunar symbol or partial solar eclipse).

See Comet Symbols....Eclipse Symbols

Independent of the issuance of E18-E23, it can be clearly demonstrated that Edward II used astronomical symbols on his coinage. In 1290, Edward II, the Crown Prince of England, was named as the count of the French province of Ponthieu, upon the death of his mother, Eleanor of Castile. But his title was contested by the Count of Alencon. The only coinage of Ponthieu during the rule of Edward prior to his becoming king has a unique design consisting of crescents and annulets.


On the 5th of September 1290, during the first year of his reign as count, an annular solar eclipse crossed directly over southern England and northern Europe, including the province of Ponthieu. Did Edward II use this celestial event as an omen for his right to authority?

The written record documents that this annular eclipse was observed. In Limoges, France, Godel (c.1320) wrote: "Also in this year [1290], on the day of Mars, before the Nativity of the Blessed Mary, in the first hour, with the moon at the 27th, there was an eclipse of the sun, with the sun itself being in Virgo, in the time of Pope Nicholas III." The partial and annular phases of the eclipse must have appeared similar to the figure shown below. Notice the similarity to the coinage design.

Now, consider the Denier au Leopard (E18-E20) and the related Denier a la Couronne (E23). An epsilon, unexplainable by Elias, is found on all coins and a crescent on some of the others. Perhaps this is a case where the epsilon is symbolic of a comet and the crescent an eclipse. An examination of the entire set of the series of Denier au Leopard Types 1 through 4 reveals that on only E18, E19, E20, E20 bis, E23 (and corresponding oboles) do both Roman and Gothic (epsilon) E's appear on the same coin. On all other issues in this series of denier where the letter E appears on a coin, its form remains consistent. Yet on these particular coins the letter E always appears as Roman in the legend but Gothic in the field.

In this case, the differentiation seems to be intentional and thus appears as strong support for the contention that the epsilon in the field is a symbol, perhaps for a comet, rather than a letter. Significant comets were taken as omens for changing kingdoms. On coins of similar design, such as the denier of Raymond III of Orange (1335-1340), a mullet is found in lieu of the epsilon which suggests that this location in the field might have been used for an astronomical symbol. If this hypothesis is accepted then it is an easy step to believe that in this particular case the crescent is also intended as an astronomical symbol, perhaps that of a partial solar eclipse.

Denier of Raymond III

In 1310, the northern coast of France was treated to an annular eclipse. Royal French territory just to the north and east of Paris, and the English held region of Ponthieu, were in the path of annularity. In Aquitaine, a very thin solar crescent would have been seen. Thus, there is strong reason to believe that the crescent found on both E20 and E20bis does indeed symbolize the eclipse of 1310.

Ponthieu Aquitaine


One can begin by accepting the arguments presented by Elias for dating E18 as the "moneta nova" of 1291. Furthermore, accept his arguments that hoard data places E20 after 1305. Since the only significant partial eclipse in Aquitaine during the reigns of either Edward I or Edward II after that date occurred in 1310, E20 must have been minted after 1310, and must therefore be attributed to Edward II. Aquitaine was in the path of annularity of another eclipse in 1321, but the 1321 eclipse does not apply to this analysis. This is clearly the case inasmuch as the design symbol would have been an annulet rather than a crescent, and there is no evidence to date this coin after 1321.

E19, with only 2 known specimens, seems to be much too rare to have been issued for approximately 2 years, and hence, cannot fill the gap vacated by E20. This rarity seems even more unusual given the relatively common occurrence of both E18 and E20. This is additional support for concluding that E19 and E20 bis are unintentional die sinker errors.

Finally, one must address E23. Here the lack of a crescent on the reverse, given the astronomical hypothesis, would appear to date it prior to the 1310 eclipse. The obverse design change could be indicative of the coronation of Edward II, hence placing it in the period of 1307-1310. This implies the following plausible sequence:

(1) Both E18 and the error E19 are now assigned to Edward I in the period of 1291-1307. If differentiating marks are needed for the mint reopening in 1305, then E18a and/or E18b (reverse legend variations) are possible candidates.

(2) E23 is an early unsuccessful issue of Edward II dating 1307-1310 (the absence of the crescent indicates a date prior to 1310).

(3) E20 and the new error E20 bis are of Edward II from 1310-1321.

This seems to be a coherent and consistent resolution to the problems associated with these issues. Moreover, with E19 disregarded as an error, E23 now seems to be a very natural fit between E18 and E20. It has the reverse of E18 and the obverse of E20 where the lion is replaced with a crown (perhaps to signify the coronation of Edward II?). One can then assume that for some reason, E23 failed to gain public acceptance, and was replaced with E20 as a reversion to the Denier au Lion. Here, the addition of a crescent not only marked a new and distinct issue, but also symbolized the 1310 eclipse and reassured the public of Edward II's divine right to rule.

Design development vs. weights and alloys

The overall appearance of E23, particularly as applies to weight and alloy, indicates that it was debased. One might argue that this is evidence that E23 postdates E20. This issue needs to be addressed.

Consider the possible coin sequences. Hoard evidence places E18 as the moneta nova of 1291 and E20 in circulation post 1311. Thus, there appears to be two plausible sequences:

(1) E18/E19 - E21a (as a transition piece without corresponding denier) - E23 - E20/E20bis;


(2) E18/E19 - E21a - E20/E20bis - E23.

From a perspective of design development, the first sequence is intuitively more reasonable. The progression is natural. Even discounting the astronomical theory that would place E20 post 1310, the crescent would be used to differentiate E20 from E18. The second sequence, however, raises a disturbing question. Since the crown differentiates between E23 and prior issues, why drop the design element of the crescent? Thus, from a design perspective, Occam's razor seems to favor the first sequence.

Now, consider weights and alloys. One should assume that whenever there was an intentional change regarding either, there should have been some change in design or privy mark to allow mint and government officials to differentiate between the old and new issues.

Known specimens of E18 and E20 indicate that they are of comparable weights. X-ray fluorescence (XRF) would be required to determine alloy definitively, but based upon the examination of representative specimens, they appear to be of similar composition. E23, on the other hand, simply appears to be debased. Moreover, the heaviest known specimen of E23 is lighter than the lightest known specimen of either E18 or E20 so, prima facie, it would seem that E18 and E20 were indeed issued at a higher standard than E23.

Thus, in order for both the first sequence and the alloy and weight analysis to be correct, one must conclude that there was a debasement in currency that was followed by a return to the earlier standard. Research by Rechenbach may resolve this issue even further.

Exchange Rates

Before addressing Rechenbach's research, let us consider the exchange rates of the period.

The Elias text indicates that the exchange rates of the 1291, 1305 and 1322-3 issues were all 5 denier per English sterling penny. Rechenbach confirms these rates. Furthermore, North states that there were no changes in the value of the sterling over that time period. Thus these three issues should be equivalent.

That should mean one of the following: (1) the apparent debasement is incorrect; (2) the coinage was debased and subsequently restored to the original standard (E18 would be the original standard, E23 debased, and E20 restored with the original design to signify restoration and the added crescent to differentiate from original); or (3) E23 is post 1324.

Rechenbach's Research

In her doctoral dissertation, Rechenbach presents further detail (pp.34-43) on the issues of Edward II based on mint records. She sets forth the following chronology, but does not assign specific coins:

(1) Earliest issue 1309-11 (2) Continued activity 1311-15 (3) Mint idle (or virtually idle) 1316-19 (4) Richard of Ellesfield's black and white coinage of 1319 (per Elias, the white coinage is probably E30, the Gros Turonus Regem) (5) Adam de Lymbergh's issue of 1322-24 (6) Debased issue as a result of the St. Sardos war in 1324 (7) Lapin Roger Tower issue of 1325 (per Elias, clearly E31, the Maille Blanche) and (8) Bordeaux revision of Tower issue struck first by Edward II in 1326 and continued by Edward III (per Elias, clearly E32, the Maille Blanche Hibernie).

Besides the chronology, Rechenbach makes two other observations from original records that are relevant to the issue at hand. First, she states that the records show that in 1311 Bernard Mandavini, the master moneyer at Bordeaux, was charged with debasing the coinage during the early issues of Edward II. Subsequent measurements demonstrated that this was not the case. This would, nonetheless, support the contention that E23 was the earliest issue of Edward II and that it was replaced with E20 to restore public confidence in the coinage. Second, she points out that the records also show that Edward II brought in Italian moneyers to operate the mint in Bordeaux. Although circumstantial, it is interesting to note that Italian moneyers introduced the crown as a design element in other European countries (e.g. The Praguer Groshcen) at the beginning of the 14th Century.

Contempory French Coinage

Medieval coinage frequently is found with similar designs among different kingdoms in nearby geographic regions. The consistency of design helped assure the general populace of acceptance of the coinage as a medium of exchange. In January, 1311, Philippe IV of France issued three new debased coins. All three coins are particularly significant to the E18-E23 analysis.

The Bourgeois fort prominently shows a large crown; one very similar to E23. This is additional support for dating E23 early in the reign of Edward II. The smaller Bourgeois simple added a new symbol to the coins of Phillipe IV: two annulets. And the Obole bourgeoise contains an annulet above the obverse cross. Recall that royal French lands were in the path of annularity during the 1310 eclipse. Thus there is further evidence for the use of astronomical symbols on issues that are closely related, both in terms of time and place, to the coins under investigation.

Bourgeois fort

Bourgeois simple

Obole bourgeoise


As indicated above, the acceptance of E19 and E20bis as die rotation errors seems to solve a variety of problems and assist in providing more coherent possible solutions to the sequencing and dating of these issues (Denier au Leopard, First Type and Denier a la Couronne). Two prominent options to sequencing then follow.

Design development (and the astronomical hypothesis) favors the sequence of E18/E19 - E21a (as a transition piece without corresponding denier) - E23 - E20/E20bis. Prima facie, weights and alloys seem to favor the sequence of E18/E19 - E21a - E20/E20bis - E23 unless, as indicated earlier, there was a debasement followed by a return to the prior standard. Given Rechenbach s chronology, E23 could thus be (1) Issue of 1309-11 (as further supported by the charges against Mandavini), (4) Ellesfield s failed black coinage of 1319 or (6) the debased St. Sardos issue of 1324. While we cannot rule out either (4) or (6) with certainty, the evidence examined above indicates that (1) is the most plausible explanation, as it can account for both design development, and weights and alloys.

Hence, when one considers the available evidence consisting of the new denier E20bis, design development, weights and alloys, celestial events, comparison to royal French coinage, and documentary evidence, it can be argued that the most consistent and plausible dating sequence and chronology for these issues can be set forth as follows:

Edward I : E18/E19 (1291-1294), E18a or E18b (1305-1307), and E21a (1305-1307)

Edward II : E23 (1309-1311) and E20/E20bis (after 1311).


Elias, E.R. Duncan; 1984; "The Anglo-Gallic Coins"; London.

Faintich, Dr. Marshall B.; 1995; "Symbolic Messengers of Medieval Man"; St. Louis.

Hewlett, Lionel Mowbray; 1920; "Anglo-Gallic coins"; London.

North, J.J.; 1991; "English Hammered Coinage", volume 2; London.

Rechenbach, Dr. Mary C.; 1975; "The Gascon Money of Edward III. A Study in Monetary History"; Ph.D. Thesis; University of Maryland.

return to main index