An Unusual Hoard of Hyperpyra of John III (1222-1256)

Written by Simon Bendall in 2018

There recently appeared on the market a hoard of some 150 or more hyperpyra of John III (1222-1256) of which so far the writer has been able to examine 78.[1] Of these most appear to be of base gold but seven appear to have no gold content at all since they are covered with a green/black patina so thick that it was difficult to identify the dies. All the coins from this hoard (PDF attached below) appear to have come from Kosovo via an Albanian vendor and were acquired by an English dealer who gave me the opportunity to study them and form the basis of this article.

These coins are of the commonest and latest variety of John III’s hyperpyra with Christ seated on a backless throne. There are five varieties (sigla) present, listed here as A – E, and, strangely, each variety seems to be struck with a single obverse die except for the second issue (B). The purity of the gold appeared to be much lower than that of the regular Nicaean coinage and to possibly deteriorate during the course of these issues with, as noted above, seven coins that have a black or very dark green patina which appear extremely base.

The coins are arranged in the attached document partly by perceived decline in their purity but mainly for the number of reverse dies for each variety. Only the last two varieties use many reverse dies and only these last two include the very basest patinated issues all of which are struck from the same dies used to strike the “gold” coins. The majority of the coins, i.e. those that are not of apparently heavily patinated “brass”, are obviously of baser gold than the regular Nicaean hyperpyra. The writer took two representative specimens that visually appeared amongst the purest and basest gold coins to a coin dealer with years of experience in scrap gold and silver. His visual examination led him to conclude that their purity was very low. An acid test revealed a very high copper content and that, in the dealer’s experience, the purity was possibly as low as two to three carats.

The dealer made enquiries and provided the writer with the name of a bullion dealer in Hatton Garden who had an XRF machine. The writer took two other coins for analysis, one which appeared to have a slightly higher gold content, no. 12, and another that appeared to be somewhat baser, no. 78. These two coins are certainly representative of all of those in the hoard. The results were quite unexpected. Neither coin contained any gold at all. The composition of no. 12 is 74% copper and 25.3% zinc and that of no. 78, 81.4% copper and 18.5% zinc. It has been pointed out to the writer that this is the same composition used to strike Roman orichalcum sestertii and dupondii which looked golden when unpatinated.[2]

There are five different varieties (sigla), all but one struck from respective single obverse dies and the reverses are struck from nine different dies with many links within each group.

The first variety (A) has three small pellets in the right obverse field above Christ’s throne (DOC 9c). There are eight “gold” specimens struck from a single obverse die of which all are struck from a single reverse die used again in all of the succeeding four issues. It should be noted that the pellets in the obverse field that form the siglum are very small, smaller than those on the regular coinage, as they also are on the following two issues.

The second variety (B) has on the obverse three pellets in the left field (DOC 9b) although these are again very small and appear in the small space between Christ’s right elbow and the throne. There are 12 “gold” specimens struck from two obverse and two reverse dies, 11 struck from one obverse die and the same reverse die used in the first issue and a single coin with an obverse die that is unique to this coin and a reverse die that is also found in issues D and E.

The third variety (C) has a small pellet above each side of Christ’s throne although the left hand one is tucked away in the same place as the three pellets in the previous variety (DOC 6c). There are three “gold” specimens struck from a single obverse die and two reverse dies, one linking two of the coins to all the other issues and one linking this issue only to issues D and E.

The fourth variety (D) has a “grenade” above Christ’s throne in the right field (DOC 12). There are 34 coins of this variety, including five base patinated ones, struck from a single obverse die and seven reverse dies, linking the issue to all except the last.

The fifth variety (E) has an annulet above Christ’s throne in the left field (DOC 6b). There are 20 coins including two base patinated ones, struck from a single obverse die and from all nine reverse dies represented in this hoard and this issue is therefore linked to all the previous issues.

The 49 regular hyperpyra of John III of this type in the DOC catalogue which are described as unclipped weigh between 3.74g (DOC 7d.1) and 5.04g (DOC 5.12). Although the coins in this hoard are basically as struck and unclipped, they are all of low weight as can be seen from the following list. As these coins purport to be hyperpyra they are, of course, the same size as the official issues and it is the lack of gold that accounts for their low weight.

In the table of the die links attached below, the obverse dies are listed as A – E with the two dies of the second issue noted as B/1 and B/2 and the reverse dies numbered 1-9.

This hoard is unusual in several respects.

XRF analysis indicates that there is no gold content.

The weights are very low, by a gram or more than the regular hyperpyra of John III, even though the coins are unworn and unclipped. The heaviest coin is lighter than the lightest unclipped specimen in the DOC collection. The absence of gold accounts for this.

There are very few dies involved with only one obverse die for each variety except for the second issue. There are few dies and many die duplicates and the die-links between the issues are greater than is usual in a group of comparable size of John III’s hyperpyra (although there were quite a number of die links in the Agrinion hoard which contained 189 hyperpyra of this type, but none between issues with differing sigla).[3]

All the coins appear to have been carefully struck with a single vertical blow unlike the usual Byzantine method of striking with two slanting blows of one or two different obverse dies which usually gives a central vertical line of overlap.[4] There is very little double striking on these coins considering the number of coins in the hoard. One coin (no. 74) has been double struck with the flan being flipped 180 degrees between strikes, something which would surely not be done by a modern forger.

The shape of the flans, which are as originally produced and have not been clipped after issue, are not individually unusual but as a group much more so since they are all roughly the same shape with none being circular but all slightly rectangular, slightly “taller” than “wider” as can be seen from the illustrations.

Both Hendy and Metcalf, to varying degrees, identify both the issues and the products of different officina by the minutiae of the design of the reverse, the number of pellets on the emperor’s collar piece and loros and the form of the pendilia of the crown.

All the coins in this hoard have two pellets on the shaft of the labarum held by the emperor which again is unusual, while the head of the labarum is missing on die D/7 and the MHP omitted between the heads of John III and the Virgin on die E/7.

Although coins with sigla three pellets in right field and one pellet in left and right fields are present in both this and the Agrinion hoard, on the coins in this hoard the pellets forming the sigla in issues A – C are much smaller. The annulet on issue E is cruder than on comparable coins in the Agrinion hoard but Metcalf notes (p. 118) that ‘there are hints that the same secret mark (siglum) can occur on coins which are very different style’. This certainly seems to be the case here.

This hoard has so many unusual features that it is necessary to consider whether all these coins are modern forgeries. This seems, at first sight, quite possible since many forgeries of later Byzantine coins have come from the Balkans in recent years. However, this seems unlikely since modern forgers have great difficulty in producing scyphates. They either cast them (Sear F91) or, if engraving dies, cannot reproduce the style successfully and make errors in the details (Sear F92).[5]

Most modern forgeries today appear to come from Bulgaria including a number of late Byzantine types but these have always comprised the flat silver Palaeologan coins which are much easier to produce, though even then the engravers cannot get the style correct.[6]

A further indication that they are not forgeries is the dark green/black patination on the basest coins and the fact that there are traces of copper (cuprite/oxide?) on a few of the ‘gold’ coins. In addition it is unlikely that a modern forger would go to the trouble of engraving so many scyphate dies in such a good, although unusual style, when their easier task of forging flat coins is so incompetent. If these are modern forgeries it might be expected that the pellets on the first three varieties would be larger as they are on the official issues and those on the left side of Christ’s throne not tucked away. Also, if these are modern forgeries why are not more dies used in issues A – C and why would a forger produce so many dies of a coin which, after all, is not uncommon and amongst the least valuable of the Byzantine gold coins? No self-respecting forger would surely produce so many dies in a such good style better than their copies of the flat Palaeologan coins and yet strike coins whose weights were so much lighter than the coins they were copying. Modern forgeries of gold coins are made of gold.

In addition, the shape of the flans is unusual. In an assemblage of this size it would be expected that many of the coins would be rounder but irregular, the usual shape of coins, but all these are the same shape and somewhat rectangular – taller than wider. On the other hand, a fellow numismatist to whom I showed the photographs immediately considered the coins false because of the similar shape of the flans which he considered had been clipped out of a sheet of brass, originally rectangular and then had the corners clipped. But why not clip the flan so that the coins appeared rounder after clipping?

Tentative Conclusions

These coins are not modern forgeries for the reasons mentioned above. Can they be Nicaean? The Nicaean hyperpyra of John III fall from about 18 to 16.75 carats with the latter figure being assigned to those considered as Latin by Pegolotti[7] and the coins published here contain no gold. If these coins were some form of official issues they would have surely been well known by now, found commonly in the Balkans and western Asia Minor.

Are they Latin? The fact that the Latins copied the hyperpyra of John III seems to be entirely due to the fact that such coins are mentioned in the notebook of an Italian merchant, Pegolotti, who describes Latin gold ‘perperi’ of 16 1/2 carats as being recognisable by having an arrangement of four or three pellets above the left side of Christ’s throne.

The writer understands that a few years ago a Bulgarian (?) numismatist, as the result of studying a large hoard, published, to his own satisfaction, what he considered were the Latin hyperpyra, apparently identifying them according to the details of the emperor’s costume, as Hendy and Metcalf had done in their arrangement of the coinage. These coins are too irregular for them to be Latin by his standards and in any case contain no gold.

These coins surely cannot be Latin for the same reasons why they cannot be Nicaean. There is contemporary evidence that immediately on conquering Constantinople the Latins began to issue coins, melting down statues to obtain the metal for a base currency and indeed the earliest issues of billon trachea attributed to the Latin empire are very common but these are much rarer and indeed several new types, unique or extremely rare, have been discovered since DOC IV was published in 1999. It seems clear that the Latin billon trachea were produced in quantity immediately after 1204 and tailed off before the accession of John III in 1222. How is it therefore that at the time the Latins were producing very few billon trachea they were able to produce a gold coin in some quantity since the hyperpyra generally attributed to the Latins appear to copy John’s later hyperpyra, those in which Hendy calls the “late” style. Indeed, are there any Latin hyperpyra at all or was the information Pegolotti recorded merely bazaar gossip? Any hyperpyra that might be attributed to the Latins do not appear to be identifiable by style.

The evidence is in favour of the antiquity of these coins and that they are not Nicaean or Latin otherwise they should be well known by now and not known from what is apparently a single hoard. Is it possible that they were struck in the general area of what is now Kosovo, whether officially or unofficially, in a single short space of time? The writer knows of no other coins like them. Do other specimens exist, struck from dies represented here? If they do and were found many years ago it would be unlikely if these coins were modern forgeries. On the other hand if these are the only known specimens of these coins, in view of the good style but the irregular details of the emperor’s dress, particularly the forms of the loros, collar piece and pendilia, together with the absence of any gold content might, suggest that these coins were the product of a superior 13th century forger. The author, however, is not categorically stating that this is so but suggests that this is a possibility and feels that it is important to publish these coins and reveal their existence.

[1] With these coins of John III there were two billon trachea of Isaac II (1185-1195, DOC 4), one clipped and one of Alexius III (1195-1204, DOC 3). Their patination is not the same as those on the base coins of John III. They may not be part of the hoard and are listed here as nos. 79, 80 and 81.

Isaac II BI Trachy

Alexius III Trachy

[2] Pinchbeck, a metal alloy which contains 89-93% copper, 7-11% zinc and which makes it look like gold.

[3] M. D. Metcalf, ‘The Agrinion Hoard’, in NC 140, 1980, pp. 113-131.

[4] S. Bendall and D. Sellwood, ‘Clipped twelfth century Byzantine electrum trachea’ in Numismatics - Witness to History, IAPN Publication 8, Wetteren 1986, pp. 97-99.

[5] D. Sear, Byzantine Coins and their Values, London 1987.

[6] S. Bendall, N Circ.

[7] F. B. Pegalotti, Pratica della mercatura (“Practice of Marketing”), 1335-1343. The 1936 edition by Allan Evans is now standard: it includes important glossaries of commodities, place names, coins and money, etc.

With thanks to Italo Vecchi (ed.).

To cite this article please use the format 'Roma Numismatics - Studia Nummorum 07022022#02

Related Articles

{{rA.name}}