The Papal-Byzantines and the Siliqua

The Problem.

The intriguing and very rare silver "Papal-Byzantine" coins* of (we presume) Rome first appear under Constans II. These are tiny coins, weighing only c.0.3g, of varying fineness, and are commonly described as "1/8th siliquae". In terms of real value (as against the theoretical siliqua of account, which was defined as 1/24th of a gold solidus) this valuation may well be correct, but I suspect that in numismatic terms it would be more appropriate to describe these types as reduced siliquae, since it seems that they are really the (more or less) direct descendants of the original siliqua coin of the fourth century.

These "Papal" types are however just the most extreme manifestation of the late silver coinage of the Roman and Byzantine empires in the west. Similar considerations presumably apply to all the other light weight late silver types of Ravenna and Carthage, which are variously described by the commentators as half, third and quarter siliquae, but which are probably all either reduced siliquae or, in the case of the fractions, reduced half-siliquae.

The question therefore is how exactly should we describe these types, and what were they really worth? To some extent of course this is a question of semantics, involving different usages of the term "siliqua", but nonetheless there are real questions to be settled here, and to answer them we need to get a proper understanding of the long and interesting history of the silver coinage of the late Roman empire, since only when this is done do we get a clear view of the varying relation between the silver (and, as it turns out, copper) coins, on the one hand, and the basic gold coinage on the other. (Of course many others before me have considered these questions in various contexts, but I haven't been able to find a concise and coherent account of the whole topic which is easily available to ordinary collectors " hence this Note).

*  Examples of most of the known types of the Papal-Byzantine issues, including many from the find reported by O'Hara and Vecchi, were offered in the important Leu-NAC sale "Arcadius to Constantine XI" of May 1993, the catalog of which has the best available pictures of these types (at 4 times actual size). Five of the coins from this sale, plus two other examples of these types, are viewable in the British Museum's online collection.

Note that many of the attributions in Sear of these types to the various imperial and papal reigns can no longer be accepted - for corrected attributions which take into account the declining silver content see the article of Morrison and Barrandon in Rev. Num. 1988, XXX, p. 149-165. This article is available on the www.persee.fr website. (Note that the Leu-NAC catalog revises a number of the Morrison-Barrandon reallocations, but not always, in my opinion, justifiably).

The "Siliqua".

The siliqua (as we now call it - its actual name is uncertain) became the primary silver coin of the 4th century, although it was by no means the only one. It was first struck by Constantine on a weight standard of 3.4g (presumably 3 scruples), probably as a successor for the argenteus, a coin of the tetrarchy known to be struck at this weight standard (1/96th of a pound), which itself was possibly meant to revive the silver denarius of Nero.

Now the value of the argenteus is uncertain - writers have variously tariffed it from 1/20 to 1/25 of an aureus (it may well have varied, and Diocletian at one stage tried to make it 1/12 of an aureus), and the intended value of Constantine's new 3 scruple coin is equally uncertain. It is generally assumed that it was tariffed as the equivalent of the "siliqua of account", a term used in financial transactions to mean a value of 1/24th of the solidus (i.e, one carat weight* of gold), although it seems that this valuation may have caused problems.

In the Roman monetary system gold was the basic reference point, so that silver coins were tariffed against the principal gold coin of the time. To accommodate the variability in the market of the value of silver bullion against gold, it was desirable that silver coins were issued with a bullion value somewhat less than their nominal value. That is, silver coins had to be overvalued to a degree sufficient to cover the normal fluctuations of the value of silver bullion.

However the relatively high weight of Constantine's 3 scruple (18 carat) siliqua meant that in fact it would often have been undervalued, or close to it, against its bullion content, as its nominal value of 1 carat of gold implied a gold to silver value ratio of 18, a figure not too different from the day to day market ratio for bullion at the time, and certainly well above the value of 12 that the mint had customarily used for coins.

This would have had two serious consequences - firstly, in periods of high silver prices, with the gold to silver ratio dropping below 18, the coin's real value as bullion would be more than its nominal value as a coin, so that it was probably liable to be hoarded rather than spent, or perhaps even melted down. Even more importantly, the government would have lost money on its production. This is very possibly the reason that the 3 scruple coin was eventually replaced by Constantius II sometime after 355 with a new coin (of the same design) with a weight of c. 2.2g, or 2 scruples; this reduced the implied gold to silver value ratio to the standard figure of 12, and gave the type a comfortable, but not excessive, overvaluation against bullion.

(An alternative scenario here might be that the original 3 scruple siliqua was tariffed at 1/16 of a solidus (of 4 scruples), consistent with a nominal gold-silver ratio of 12. But then the change in weight would have been both unnecessary and undesirable, as it would have resulted in an unacceptably overvalued siliqua. In any case, whatever the reason for the weight change it seems reasonable to assume that the 2 scruple coin was meant to be worth 1/24 of a solidus).

*  We need to distinguish the siliqua coin from the siliqua weight. "Carat", in Greek keration, means a carob seed - "siliqua" is the Latin equivalent. In weight terms the siliqua/carat was 1/24 of the solidus/nomisma weight, the latter being 1/6 of an ounce. Since a scruple was 1/24 of an ounce we see that the scruple was therefore 6 siliquae/carats in weight terms. The "siliqua" coin of 2 scruples was then 12 siliquae/carats in weight.

The Eastern Empire.

In Constantinople at least the weight standard of the siliqua was maintained at 2 gm or so until the mid 5th century, when it dropped significantly for a time, to c.1.4g under Theodosius II, and 1.3g under Leo I. The weight was then, it seems, restored to the previous standard by the time of Zeno and Anastasius (search on zeno siliqua const 491) and this level was maintained, albeit with very small issues, until the early 7th century when the siliqua was replaced by the hexagram, a new silver coin with the weight of 3 siliquae, i.e, 6 scruples.

In fact from the later 5th century onwards the siliqua plays little part in the currency of the east, and its place in the monetary system was taken by the new copper coinage of Anastasius although we also see some local(?) anonymous silver types with weights of c.0.9g, possibly tariffed as half-siliqua. After the new hexagram was introduced the siliqua continued for a period as a ceremonial type, together with its double, the miliaresion.

(Exactly what role the not very common anonymous types played in the monetary system of the 6th century is a matter of conjecture - they, and their similar predecessors of the 4th century, have been variously described as commemoratives, local municipal issues and public donatives ("throw money"). The basic problem with these anonymous issues is their weight, which is generally too light for a half-siliqua in the official system, and too heavy for a third-siliqua. Most likely these are, I would guess, tokens of debased silver used for account clearance in the markets of the major cities, with a value that could have been anything from 1/4 to 1/2 siliqua).

Note that there is also a rare silver type (presumably from an eastern mint) in the name of Anastasius, weighing about 1 gm, with a large T on the reverse. This has sometimes been taken, because of the T, to be a 300 nummi type, which would presumably make it a lightweight siliqua (see below), but I suspect that it is really just a variety of the anonymous "half" siliqua Bendall 11, where the T indicates, perhaps, some mint.

The Western Empire.

In the declining western empire the siliqua's ultimate fate was rather different; whereas in the east the silver coins by and large retained their weight, and hence their effective value, in the west they appear to have slowly and steadily degenerated into tiny coins whose real value we can now know only approximately.

In the earlier and mid 4th century the siliqua was produced in quantity in the western empire and obviously played an active role in the monetary system there. Through the latter part of the century production of the siliqua, now mostly standardised with a seated Roma on the reverse, continued at reasonable levels, but while its theoretical weight (as indicated by the heaviest coins) was maintained at around 2.2g, a little before the end of the 5th century the average weight of the circulating coins started to drop, reducing to c.1.5g by the end of the reign of Honorius. It is sometimes said that this was due to systematic clipping that began in Britain after the Roman withdrawal, but clipping appears to be primarily a British phenomenon, while most western mints (outside Italy) seem to have been striking siliquae on underweight flans anyway from the time of Arcadius onward. Lighter examples of the Roma seated left issues of this period are described in some sources as "half-siliquae", but given the varying issue weights of these types they (and the votive types on the same weight scale) were presumably actually all supposed to be siliquae, even though they are often not much heavier than the Victory turning to left types, which do seem to be real half-siliquae, and which, for some reason, did not lose any weight.

After Honorius central authority collapsed in Gaul, and official production of silver coinage largely ceased in the west, although siliquae in the old Roma seated design were still occasionally struck in Italy. However, after Valentinian III some new coins in different designs were also issued, mostly at weights of around 1 gm. These new types, which continued until the introduction of the Ostrogothic coins, are usually described as "half-siliquae", which they may well have been (there are also some very rare late seated Roma types of Julius Nepos weighing c.2g, which seem to be siliquae on the same weight scale). It seems that in Italy in the later 5th century an effort may have been made to maintain the old weight scale of the siliqua, although only with nominal issues.

At the end of the reign of Honorius the Vandals overran North Africa, and in Carthage they initially issued silver coins in two sizes copying the siliquae and half-siliquae of Honorius, with an average weight of c.1.6g for the siliqua, and half that for the half-siliquae; these imitative types were often struck on undersized flans, so that they looked much like the clipped imperial coins circulating at the time. These types were issued in considerable quantities, and probably over some time, but ultimately the Vandals developed their own distinctive and quite different silver coinage, involving types of 100, 50 and 25 "denarii" (i.e, "denarii communes", the common unit of account, of perhaps 2 1/2, or more likely, 5 nummi). The weight standard of the initial series of these new types was based on a 100 denarius coin of a little more than 2g, i.e, roughly the weight of the old siliqua, and, while the weights of the later series (consisting mostly of 50 denarius types) varied a little over time, on the whole they did not vary too much from this basic standard.

In the late 5th century the Ostroths in Italy began regular and substantial issues of their own silver coins which extended well into the 6th century. These types come in two sizes, weighing c.1.4 and 0.7g, which are generally described in the books as "half-siliquae" and "quarter-siliquae". However on the face of it they would seem to be too heavy for these denominations, at least on the original Roman scale, so perhaps they were actually meant to be siliquae and half-siliquae on a weight scale derived from that of the common imitative types of the Vandals, which must have been well known in Italy in this period, and which in turn recalled the reduced weight and clipped imperial coins in Gaul from the earlier 5th century. (It's noticable that neither of the new Ostrogothic denominations matches the weights of the official Italian types weighing c.1g that immediately preceded them in Italy, which, as we have seen, may well have been issued as full weight half-siliquae, although only in very small numbers). It's worth noting here that the Ostrogothic types exhibit a distinct change in the style of the obverse bust from that of the earlier Roman types (the last of which were issued by Odoacer in the name of the eastern emperor Zeno). The bust is now shown from the rear, rather than the front, and this change would persist after the return of the Byzantines.

In the 6th century the Byzantines regained control of much of Italy for a time and resumed production of the siliqua in Ravenna (Rome in the older books). Under Justinian I the weight scale was initially c.1.4g, similar to that of the Ostrogothic issues, but it then dropped to 1.05g (1 scruple perhaps?). In both of these weight series some of the types are labelled as 250 and 125 (meaning either nummi, or, perhaps, denarii of 2 nummi), and as the two main denominations here would seem to have been the direct successors to the Ostrogothic types they were presumably meant to be siliquae and half-siliquae (although other interpretations are possible, as discussed below). There are also types (Sear 322), at least some of which are apparently official, with no value label and a weight of c.0.35g; these could in principle be quarter-siliquae of the heavy series, but another, and I think more likely, possibility is that they are half-siliquae of a late extra low weight series of Justinian at Ravenna (struck on the same weight scale as the subsequent issues of Justin II with similar designs), in which case the unlabelled 0.7g type (S.320) would presumably have to be the corresponding siliqua. (S.320 even manifests the same two main styles - neat and loose - that are found in S.412 and 413 of Justin II). S.318 is then presumably a siliqua of the light (second) series, not a half-siliqua.

(For a summary of the various issues of Justinian at Ravenna see the Appendix below. Note that the position adopted here in respect of these types is basically the same as that taken in the Leu-NAC sale of May 1993 referred to in the introduction - cf. lots 487-491 - although there the question of siliqua versus half-siliqua is judiciously avoided).

Under Justin II the weight scale at Ravenna dropped again (or, as just suggested, the late extra-light standard under Justinian was maintained), since the coins, some again labelled with the two denominations, now weigh c.0.7g and 0.4g, and as well there is evidence (in the form of even lighter weight types such as the "starred" version of the 250 unit type S.411, Ran. 417) that a second series was issued by Justin II with a siliqua weight of c.0.5g. There are also crude versions of some types which are possibly Lombard copies - whether the relatively common type S.413 should be included in the latter is not clear. (Note that the undenominated types of Justin II S.412-15 are easily confused with similar extra-light issues of Justinian S.320 & 322 with the same weights).

After Justin II the silver coins of Ravenna are generally rare, except for a partial revival under Heraclius. Under Tiberius II we are left with only the prime denomination, now weighing about 0.5 gm. Under Maurice the weight drops again to around 0.4g (1/3 scruple?), and this carries through to Constantine IV and Tiberius III. (I know there is a rare "siliqua" of Heraclius attributed to Ravenna (S.904A) weighing 1.6 gm, but this appears out of place here and is perhaps actually the issue of an eastern mint).

In Carthage under the Byzantines we see a basically similar although not identical pattern (and note that here, as in the east, the general style remained Roman). To start with the weights are more or less constant - under Justinian we have silver types weighing 1.2 and 0.6g, while under Justin II the primary coin weighs c.1.1g. Under Tiberius II we again have coins weighing c.1.1g and 0.6g. These types could have been siliquae and half-siliquae, as was (perhaps) the case at Ravenna, but half and quarter-siliquae are possibly more likely here, as successors to the Vandalic 50 and 25 denarius types with similar weights. Under Maurice and his son Theodosius the weights start to decline, to 0.95g for the main type, together with rare fractions of c.0.5g, and by the time of Heraclius to 0.7g with a fraction (apparently unique) of 0.35g. Under Constans II all coins are from a single primary denomination which averages 0.55g.

The Value of the "Siliqua".

The obvious question now is what were these variously sized types actually worth? In the late 4th and early 5th century the circulating clipped (and reduced weight) coins had presumably been originally issued as "siliquae" (i.e, with the value of the siliqua of account), and presumably they also generally traded as such in the market, at least to begin with, otherwise there would have been no point in clipping them. However, as King has suggested, mixed hoards suggest that in later times these coins were likely treated as bullion in Britain.

In the 6th century we have coins issued at Ravenna in the name of Justinian and Justin II on weight scales of 1.4g, 1.05g and 0.7g, all labelled 250, but 250 what? In the earliest stages of his monetary reforms Anastasius issued "folles" of 40 nummi weighing c. 9 gm, with the solidus worth 420 folles = 16800 nummi. Anastasius then doubled the weight of the follis, so that the solidus was now worth 210 (new) folles = 8400 nummi, and ultimately in 538/9 Justinian increased the weight of Anastasius's follis slightly and hence reduced the number of nummi in the solidus to 7200 (180 folles), matching the figure in the time of Valentinian III a hundred years earlier. This last figure is consistent with a siliqua (or whatever the primary silver coin was actually called) of 300 nummi tariffed at 24 to the solidus. But while this presumably applied in the east, in the west the 250 nummi pieces could perhaps also have been tariffed as siliquae, with 24 of these types, and hence 6000 (western) nummi, or 150 (western) folles, to the solidus, at least for the heaviest series. In this scenario then the western folles and nummi are worth 180/150 = 6/5 times their eastern counterparts.

But there are other possible scenarios. Indeed, as we have noted the standard theory in the literature treats the 1.4 gm coins of both the Ostrogoths and Byzantines as half-siliquae (rather than reduced siliquae), with, in the latter case at least, 48 half-siliquae of 250 nummi, or 12000 nummi (= 300 folles), per solidus.

However, there are difficulties with such a rerating, as a half-siliqua of 1.4g rather (rather than 1 scruple = 1.14g for the old official coin of Constantius) would have raised the nominal gold-silver ratio from 12 to c.15, thus partially reversing Constantius's original reform, and significantly increasing the treasury's costs of producing the coinage without any apparent necessity to do so. Also, if the Ostrogoths had really wanted to reform the silver coinage the most obvious way to do it would have been to restore the old official weight system (as the Vandals ultimately seem to have done, at least approximately, and as had in any case apparently also been done in late pre-Ostrogothic Italy itself), meaning a half-siliqua of 1.14g rather than 1.4g. This would also of course have been a cheaper option for the government in terms of silver.

However, in the event we see that the weights were not significantly changed from those of the older Vandalic imitative types, which prima facie supports the idea that the Ostrogoths, and the Byzantines who followed them, preferred to revert to the reduced siliquae of the early to mid 5th century, as suggested earlier.

Alternatively, we could possibly, as Hendy suggested, read "250" as 250 denarii = 500 nummi. Assuming these types were still rated as siliquae of account, this would effectively halve the value of the nummus, resulting in a solidus of 24 siliquae = 6000 denarii, equal, once again, to 12000 (reduced) nummi. However this would mean using different units for the silver and bronze types, ie, while 1 unit presumably means 1 nummus as usual on bronze coins, it must be read as 1 denarius = 2 nummi on the silver types, and there is no real indication that this was actually the case, as there was with the (different) mixed units of the later Vandalic coinage.

Nonetheless, one point worth considering here is the significantly lower weight scale of the western folles of this period as compared with that of their eastern counterparts - could this perhaps reflect a reduced value of the western nummus along the lines suggested by Hendy, which would theoretically now be valued at 3/5 of the eastern version?

In any event, whatever the absolute value of the silver types may have been, under Justinian we soon see the resumption of the classic pattern of devaluation, with declining weights of the issued types followed no doubt by a declining real value against the solidus in the market. Assuming the basic coin was a siliqua, the light series types of Justinian would then have been exchanged at 32 to the solidus, and the extra-light series at 48, so that the primary extra-light type would have had a value of 1/2 of the siliqua of account, and so on (although we have no direct evidence to confirm these values). By the time of Constans II, then, the basic coin would likely have been been worth 1/4 of the siliqua of account. (These values can be halved on the assumption of a basic half-siliqua to start with).

Given the decline in the value of the nummus against the solidus implicit in the declining weight of the 250 unit coins, it is not surprising that we also see a decline in the weights of the copper coins in the west, and the eventual disappearance of the smaller denominations, both indicating a substantial overall drop in the value of the copper types against the solidus, with the follis possibly reaching something like 1200 or more to the solidus, as against its original value of 150 (or 2400, assuming originally 300 to the solidus).

In the east the weights of the copper types were generally higher than in the west, although eventually they declined as well, particularly under Constans II and later rulers. Once again the disappearance of the smaller denominations shows that this decline in weight reflects a real decline in the value of the nummus, and hence of the follis, against the solidus (and in this case also against the silver coins which, as we have seen, maintained their weight and hence presumably their value in the east).

The Papal Types.

Finally we return to the "Papal-Byzantine" silver types of Rome. As noted at the beginning, these first appear under Constans II, weighing about 0.3g, and they extend on basically this weight scale through to the time of Pope Zachary early in the reign of Constantine V, although with a steadily declining silver content. Were these types regular coins? Or were they really "ceremonial" types of some type (like the later papal enthronement medals), not meant to play any significant role in daily commerce, or perhaps more plausibly, were they charity tokens distributed to the needy? It's hard to say definitely, but the fact that they were first issued at around the reduced weight of the other silver types in Italy at the time suggests that these types were indeed real coins, perhaps issued by the pope as prefect of Rome, or at least were meant (if they were tokens) to be valued as real coins. Hence I suggest that they should be regarded, like the other silver types of that period, as highly reduced siliquae, or perhaps reduced half-siliquae if we take the primary ostrogothic coins to be half-siliquae.

In terms of value we have taken Constantius's siliqua of 2 scruples to be equal to a siliqua of account, so at c.0.3g the papal siliquae would have had a nominal value of c.1/8th of the siliqua of account, and possibly, allowing for the debased alloy, even less.

It should be noted by the way that the alternative rating of the papal siliquae as "30 nummi" (as in Sear) seems to mean standard nummi of the mid 5th century, before the decline of the silver and bronze coins in the west. In fact the papal types were probably nominally rated as equal to 250 of the reduced nummi of their own time, although once again in practice this figure may have been reduced by the debasement of the alloy. (As noted in the Leu-NAC sale catalog, the reverse of one particular type can possibly be read as both CON(s)T(antine) and C/N, i.e, 250, as with the earlier "quarter siliquae" S.1135-1136A of Constans II at Ravenna).

In other words the papal siliquae are clearly not the equivalent of the 30 nummi "billon" (but more likely bronze) XXX types of Rome in the later period, or anything like it.

Note that most examples of the bronze XXX types generally weigh c. 0.5 - 1.5 gm, and are on essentially the same highly reduced weight scale as the last folles of Ravenna, namely those issued in the names of Leo III and (perhaps) Constantine V. This supports the general assumption that they date from much the same period as the papal siliquae, although it should be noted that stylistically there is little to match most of the bronze types to particular issues of the papal siliquae, and in fact we can't even be sure that they were struck on same authority as the silver types. (Examples of the XXX types can be be found in the Leu-NAC sale and also the Berk-England sale of Dec. 1989).

Ross Glanfield

January 2012.

Appendix: The issues of Justinian at Ravenna.

The issues of Justinian at Ravenna can now be summarised as below, on the assumption that the primary coin is a siliqua. Note that all "siliquae" here should be understood as having a value of 250 units, taken here to be (western) nummi.

Heavy Types, Siliquae of 1.4g:

  • S.313:  CN (= 250 nummi) reverse - siliqua.
  • S.315:  PKE (= 125 nummi) reverse - siliqua.
  • S.317:  PK  (= 120 nummi) reverse - 3 folles (1/50th of a solidus).
  • S.319:  XP monogram reverse - siliqua.

Lighter Types, Siliquae of 1.05g.

  • S.314:  CN reverse - siliqua.
  • S.316:  PKE reverse - siliqua.
  • S.318:  Cross on globus reverse - siliqua.
  • S.321:  Large star reverse - siliqua.

Extra-light Types, Siliqua of 0.7g (as Justin II).

  • S.320:  P-headed cross on globus between stars on reverse - siliqua.
  • S.322:  P-headed cross on globus (no stars) on reverse - siliqua.

S.320 divides mainly into two basic types - one with a neat dotted style, as on the earlier types, with a smallish globus below the cross, usually above a flat (stepped?) base, and another with a loose linear style with a larger globus below a cross with a forked base. With minor variations the same two styles are found on the later issues S.412 and 413 of Justin II (and also on the heavier versions of the CN type S.411). Note that although these styles differ, they are both generally similar to better examples of the preceding Ostrogothic types (and rather different from the earlier issues of Odoacer which exhibit the style of the last Roman types). Do these two styles suggest two separate mints for S.320, one Byzantine and one Lombard? It's hard to say definitely, but in the end only the original dotted style survives into later reigns at Ravenna.

Exactly where the rare type S.321 (Large star reverse - copying the scarce issue MEC 117 of Theoderic in the name of Anastasius) fits is not clear- its weight is somewhat uncertain but it is taken here to be a 1/2 siliqua companion of S. 318 in the light series. S.322A, with a cross potent on a globus, weighs c. 0.35g and again it is not clear where it fits - its loose style suggests it is perhaps an imitative type.

Latest revisions:

  4 Feb. '12:  Leu-NAC Sale of May 1993 noted. 
22 Mar. '12: Style change under Ostrogoths noted.  
23 Sept. '14:  Siliqua versus half-siliqua question reconsidered.  
  1 Dec. '14:  And further revised. 
  3 Dec. '23:  General discussion revised.

[Home] [Articles] [Notes] [Data]