The Coins of Croesus (Kroisos).
The earliest issue in Croesus's new bimetallic series was very likely Berk's (or rather Naster's) "Prototype" gold typea, a "heavy" stater of c.10.7 gm. The basis of
this weight scale is unclear, but we note that it is close to but clearly less than twice the weight of the later Persian silver siglos, i.e, 2 x 5.55 = 11.1 gm, which is in turn related to the weight of the
Babylonian/Persian shekel (see Appendix for details of the weight scales - at this stage of course Persia had issued no actual coinage). This type is quite scarce, and this, together with the distinctive and lively
style in which the animals are represented, suggests strongly that this was indeed the first issue of Croesus. This conclusion is reinforced, as Berk noted, by the fact that this type shows a dot on the lion's
forehead, a feature that seems to connect it to the "Lydian" electrum lion head types.
Berk also nominates some prototype silver issues (which are not shown here). Personally, I don't find these at all convincing, as they have no real similarity to the gold prototypes - there is no dot
on the lion's head to start with, as Berk also noted - and in any case one of the nominated types is struck on a flattened flan, so that overall these types seem to be simply slightly odd issues of a later date.
Perhaps no silver coins were issued at the earliest stage because of the novelty of the type as compared with electrum and gold.*
After the prototype we see a range of types in the standard style - heavy and light gold staters, silver staters and half-staters, and fractional types, in both gold and silver, based on 1/3rd
The light gold staters were struck at c. 8.05 gm (not 8.15 gm, as is often stated), somewhat below the Babylonian/Persian shekel weight of 8.33 gm, and as noted in the Appendix, were presumably equal
to 10 silver staters of 10.7 gm. The curious mix of denominations here recalls the early stages of the Ionian electrum coinage, and just how all the various types fitted into the monetary system is an interesting
question, but not one which we can pursue in detail here, except to surmise that it may be that different types circulated in different areas, or perhaps served different markets.
Apart from the prototypes, the Croeseids have been divided and classified in various ways, principally on the basis of style, but also taking into account weight standards (i.e, the heavy and light
gold stater series) and known die and punch matches.
In 1961 Nasterb saw four main groups, which he called Massive, Forceful (or, oddly, "nervous", if that's the correct translation), Transitional and Fully Stylised. Nimchuk, in her article of 2000c, defined 6 types, dividing Naster's Massive group into Types A and B, calling the Forceful group Type C, covering the Transitional group with the very similar Types D and E, and denoting the Fully
Stylised group as Type F.
Personally I think we can divide the Croeseids into three main groups, designated I, II & III, which I call Massive, Refined and Evolved (or, if you prefer, Stylised). These groups are defined
below, and a chart of images of characteristic types can be found on the "Croeseid Types" page.
* The examples of the silver "prototypes" that I have seen offered in trade seem to be basically ordinary types in a somewhat lighter than usual style (which in any case do not match Berk's
The fractions are mostly 1/3rd staters and binary sub-multiples thereof but there are also a (very) few examples of silver fractions which could possibly be construed as 1/8, 1/16 and 1/32 staters (see e.g, CNGe354-168 and CNGe277-72). The place of these rare types in the monetary system is unclear and they are not considered further here.
I - Massive.
The massive period includes all the types mentioned earlier - heavy and light gold staters, silver staters and half-staters, and fractional types* in both gold and silver.
To adapt Naster, the types here have large, powerful and stocky forms, with fine details in the lion's mane. Nimchuk divides this group into a "heavier" (in style) subgroup Type A and a "lighter"
subgroup Type B, but in my opinion this is an oversimplification, as it is often difficult to decide to which type a given coin should be assigned, and in any case there isn't much difference between the two types
to start with. In fact we actually seem to have a single group of issues with a mix of similar styles, ranging from more or less Type A to more or less Type B, rather than two clearly differentiated series of
Nonetheless, with the gold types we can, working just on the basis of the heaviness of the lion's shoulder and forearm, effect (with some effort) some sort of stylistic division of these types into
two subgroups which seem to generally agree with Nimchuk's idea of Types A and B (uncertain intermediate examples can be classed as AB). Whether this approach yields a meaningful division into separate issues though
is uncertain, particularly as it's noticeable that the two subgroups often share reverse punches**.
With the silver types we can adopt a similar approach, although here we tend to get a range of medium to heavy set types which can be classed as Type A's, plus a smaller group of somewhat lighter set
types which can be called Type B's^. However, here there may be some significance in the different styles, as the Type B silver staters probably date from a significantly later period than most of the Type
A's (see below). Ultimately though the difference is a matter of judgement, and in the end it may prove more meaningful to subdivide these types according to the style of the bull's neck folds, rather than general
It is noticeable that while the gold types of this group are generally struck on unflattened (or perhaps semi-flattened) flans, the silver staters (but not the half-staters^^) are often struck on semi-flattened or flattened flans, suggesting (contra Nimchuk?) that production of the silver staters in the style of this group continued for some time after the gold issues had
More generally, the fact that flattened flans start to appear during the period of this group (on the silver staters) and the next (on the gold types) would seem to indicate that these earlier
Croeseid types were issued in parallel with the later issues of the Lydian lion-head electrum types of Weidauer Group XVI, which were also flattened (and which may, or may not, have been struck at Sardis).
One question of some interest here is whether production of the heavy and light series of gold types overlapped in time, as they seem to have circulated together. It's possible, but if there was an
overlap it was probably only limited. It seems that the two series included both Type A and B style coins, but as we have seen this may simply mean that there was a single group of slightly varying styles throughout
this period. More significant is the fact that only one of the two punch sets of the heavy series is found in the light series, and as well we note that there is only one weight standard for the silver staters, i.e,
there are no "heavy" silver staters to match to the heavy gold types, suggesting (perhaps) that the silver types only appear with the light gold series, following the heavy series. (Although it's possible I suppose
that Croesus struck his earliest gold and silver staters at the same weight of 10.7 gm, but still with a nominal value ratio of 10 to 1, and then lowered the weight of the gold types - see also the Appendix for
further ideas on the question of the heavy staters. In the chart I hedge my bets on the question of an overlap, so you can make up your own minds).
Judging from the number of different punches involved (and the number of coins appearing in the market) fractional types made up a considerable part of the coinage in the massive period. This is not perhaps the impression that we gain from Nimchuk, but the material that she worked from (mostly from museum collections) seems to have been relatively deficient in fractions. Fractional silver types can be hard to classify in detail in Nimchuk's terms, but they are found with both unflattened and (less commonly) flattened flans, suggesting that they were issued throughout the massive period, and beyond it.
In my opinion the heavy gold staters can mostly be described as Nimchuk Type A, with some Type B's, while the light staters are spread evenly between Type A and B, but I doubt that this means much. In any case there is no clear separation of the two nominal types between the heavy and light series.
Nimchuk doesn't recognise Type B silver staters, but her example Paris 537 (accessible online as BNF cb41764720k), which is listed as a Type B heavy gold stater (with incorrect punches) in her article of 2000, is in fact a silver stater. This coin is shown as No. 10 on p.42 of the article, and also as the cover coin of the journal; for similar coins with the same punches see Triton VI-408 and NAC 96-1104. Coins like these, although not noticed by Nimchuk, are not uncommon and can reasonably be classed as Type B, although some examples are offered as "Prototype" silver issues. Two examples are shown here (under "flattened flans"); as noted on the chart, and explained later in this section, it would seem that these types postdate the Group II issues. It would appear that the silver staters are more complex than Nimchuk realised.
Nimchuk describes the silver half-staters of this period as Type B, but to me Type A seems more appropriate in most cases. Nimchuk maintains that the half-staters were all struck after the silver staters, but the fact that most of these types are struck on unflattened flans suggests that they were contemporary with the earlier full staters, although I am not aware of any shared dies or punches. On the whole the silver half-staters seem to have played a relatively minor role in the coinage of the massive period. Incidentally I am not convinced by Nimchuk's identification of her obverse type kk on the half-staters with the obverse type g on the light gold staters (pp.14 & 32) - cf. ANS 1955.54.397 and BM RPK,p146B.1.Sam.
# It is worth noting that Type A silver staters are struck both on unflattened (or semi-flattened) and flattened flans, while the Type B's mostly feature flattened flans.
II - Refined.
This is essentially Naster's second group, and I use the term to cover types like the gold staters that made up the bulk of the Sardis hoard of 1922, which are somewhat lighter and finer in style than
the massive types, and where the lion's mane is shown in dots rather than the usual dashes or ovals. As well, with the larger denominations at least the bull is generally somewhat smaller than the lion, as opposed
to the massive group, where the lion and bull are basically the same size.
This is a well defined issue, presumably limited in time, involving a relatively small number of obverse dies and reverse punches. It is confined mainly to light gold staters, some of which are punch
linked to some light gold staters of Group I, although there are also a few silver half-staters, some from the same obverse dies as the gold staters*.
As well there are light gold fractions (and a few silver fractions) with dotted manes which possibly fit in this group, and these are included in the chart here.
Some of the gold staters of this group are struck on flattened flans, and this, together with the punch links, suggests that this group followed closely the gold types of the previous group, and that
it was contemporary with the last issues of the Lydian lion-head electrum types.
It is noticeable that there appear to be no full silver staters in the Refined style, although as noted earlier plenty of Group I style silver staters (and some silver fractions) were struck on
flattened flans, presumably dating from after the unflattened Group II gold issues. This suggests that the Group II issues derived from a particular workshop devoted largely (but not entirely) to gold types, while silver staters and fractions continued to be produced by a separate workshop using the earlier style. For an example of a later fraction in mixed style see the 1/3 stater from the Money Museum (shown here), struck on a flattened flan and showing a large bull's head and a dotted lion's mane.
* It should also be noted that there are some odd issues stylistically intermediate between Types B and C in both gold and silver, with smallish bull's heads but dashed lion's manes. For an
example see the heavy gold type illustrated as Type C in Nimchuk (No. 15, p.42), which is really closer to Type B than anything else.
III - Evolved.
This group covers Naster's Transitional and Stylised types. Here the obverse details are coarser (although not cruder), with larger and hence fewer dashes in the lion's mane. As well the heads are
more stylised, with the elements segmented, and in contrast to the previous groups the ruff of the lion is normally more angled, often to c.45 degrees. Finally the lion's tongue is mostly shown as little more than
one or two dots. The types in this group seem to be largely confined to light gold staters and silver half-staters, as at this stage the 1/3rd staters and their fractions have largely disappeared, although I have
included a couple of possible examples of the smaller types (in silver only) in the chart as well. (Whether these last types really belong in this period is hard to say, although note the flattened flans on the
As stated earlier Nimchuk's type E corresponds to Naster's Transitional group, and the (somewhat) more simplified type F to the Stylised group. Nimchuk's type D is a bit of a problem, as it seems to
me that Nimchuk's exemplar is really not much different to Type E, and hence it is not included here. Instead I have here introduced a type stylistically intermediate between Nimchuk's B and E, or to put it another
way, a somewhat simplified Type B with a sloping ruff. This type has some evolved features but the heavy style and unflattened flans suggest it probably belongs with Group I rather than Group III, and it is shown as
the "evolving" Type BE on the chart.
If you like you can subdivide this group into subgroups IIIe and IIIf along the general lines of Nimchuk's classification of Types E and F. However, it's difficult to define a clear dividing line
between these two types on the basis of the Nimchuk's somewhat vague notion of style, and so I have redefined the types on a more objective basis. Type E is here taken to mean types where the bull's neck folds are
re-curved (i.e, S shaped, as on Type C), and Type F to mean types where they are simple arcs. This definition seems to correspond fairly closely to Nimchuk's stylistic separation anyway, although we now have
significant numbers of gold Type E's, a class which Nimchuk didn't recognise, but which we should surely expect. Note also that with the revised definitions some Type E gold staters and silver half-staters share
common dies. All of this supports the validity of the new definitions.
Whether the two types (however defined) correspond to two separate issues (or groups of issues) or not is not clear however, and until we have more data on die and/or punch matches it will be
difficult to convincingly separate the group into specific issues*.
Dating this group is another problem, but given the simplified style of the types, and the absence of heavy gold staters, it seems reasonable to assume that this is the latest of the three groups; in
fact some writers think, very likely correctly, that these types were struck either wholly or partly under Persian rule, and there is some hoard evidence supporting this idea. It should also be noted that flattened
flans are largely the norm for both the gold and silver coins of this group, again confirming a relatively late date.
* Nimchuk found that some (of her) Type E and F silver half-staters shared reverse punches, but with the revised definitions here I have so far found no punches common to both of these types
(because, in some cases, Nimchuk's Type E's become Type F's under the new definitions). This is consistent with the idea that the revised types do in fact derive from two different groups of issues.
Finally, we recognise that there are a number of odd looking types that don't really fit stylistically into any of the major groups above. Some of these may be genuine ancient issues, but others may
not, and none of them are shown in the chart here. (For two particularly interesting examples, see the silver staters BNF 1966.453.2793 & 1966.453.2794).
a: H.J. Berk, "The Coinage of Croesus: Another Look," SAN XX, 1 (1997), p. 14-15.
b: P. Naster, "Remarques charactéroscopiques et technologiques au sujet des créséides", in Congresso Internationale di Numismatica, Rome (1961), v. 2, p. 25-37.
c: C.L. Nimchuk, "The Lion and Bull Coinage of Croesus", Journal of the Classical and Medieval Numismatic Society 2nd ser., 1 (2000), p. 5-44.
Appendix - The Persian and Lydian weight standards.
In the 6th century the Persian shekel (as a weight) was 1/60th of the Babylonian/Persian mina. Since we know from surviving weights that the mina of this period weighed close to 500 gm*, we find that the weight of the shekel was c.8.33 gm, which is in fact exactly the average weight as found of Darius's gold daric coin from the latter part of the century. The Persian silver "siglos" (shekel) coin of the late 6th century was evidently meant to be worth 1/20th of a standard shekel of gold - its weight of 5.55 gm was 2/3 that of a shekel, corresponding to a nominal gold to silver value ratio of 40/3 (slightly higher than the figure of 13 reported by Herodotus for the time of Croesus).
Croesus' somewhat earlier (light) gold and silver coins were evidently related along the same lines, although on a slightly lighter weight scale. Thus Croesus' silver stater (double-siglos) coin was
meant to be worth 1/10th the value of his light gold stater, as its weight of c.10.7 gm was 4/3 times the weight of the light gold stater of 8.05 gm. These weights were probably carried into the reign of Cyrus, but
in the time of Darius the weight of the siglos was revised slightly upwards from 5.35 gm to 5.55 gm, bringing it into line with the weight of the new daric coin.
Thus it would seem that Croesus issued his light gold coins (and hence the linked silver types) at slightly below the Persian weight standard. Just what the basis of Croesus' weight standard was is
unclear. Was the light stater meant to be worth a Persian shekel weight of gold, so that the difference from the Persian standard simply reflected seigniorage of around 4%? Perhaps, although this seems a somewhat
high level of seigniorage for gold coins, so, given that Persia had no coinage at this stage, a more likely possibility is that Croesus' standard was designed to be consistent with that of the contemporary Lydian
electrum coinage, since the value of Croesus' light gold stater matches reasonably well the (actual) value of the contemporary Lydian standard electrum stater (i.e, three trites) of c.14.1 gm, if we accept current
thinking that the gold content of the latter is c.54%.**
(The figure of 54% for the gold content might seem rather arbitrary at first sight, but this only reinforces the idea that it was chosen deliberately. In practice it could reflect a nominal
gold/silver weight ratio in the electrum of perhaps 6/5 or 7/6, or possibly even 13/11, i.e, 13 carat fineness in modern terms - remember that 24 was a common multiplier/divisor in ancient measurement systems.
Alternatively, O.N. Melnikov prefers a ratio of 9/7, i.e, 56%, for the Milesian staters (which he assumes were struck as "Phoenician royal staters", i.e, stater/didrachms on the full Euboic weight standard), as this
divides the stater into 16 units, with the unit supposedly corresponding to a "Babylonian" gerah/carat (with the latter taken to be 1/24 of an assumed Babylonian silver scale double-shekel of 21.8 gm); whatever the
merits of these rather convoluted ideas, the value for the ratio seems a little high - hopefully time and more measurements will resolve the question).
One question still remains to be answered - why did Croesus issue heavy gold staters of 10.7 gm to begin with? One theory is that this weight resulted from an official position that coin electrum was
supposed to be rated against silver at the traditional ratio of 10:1, as then an electrum stater of 14.1 gm (actually three trites) would be equivalent to 141 gms of silver and hence to a gold stater of 141 x 3/40 =
10.6 gm, almost (but not quite?) the actual weight of the heavy staters. However, this valuation of electrum implies a gold content of 73% (or 75%, depending on whether the value of the silver is included or not),
which is a reasonable value for natural electrum, but far higher than the actual value for coined electrum in most cases, so that redeeming electrum for gold at this rate would have been very expensive for the
state. And so (the theory goes) the light gold stater was then introduced to match the actual value of three electrum trites, i.e, the (virtual) Lydian electrum stater, which would have saved the state a lot of
Well, maybe, but we might perhaps doubt that this inflated exchange rate really ever was (or was still in Croesus' time) the official position in Lydia , and maybe the heavy gold stater was actually
meant to be traded for four electrum trites (rather than three), since the real values seem to match pretty well. But if so why was it then replaced by the light stater? Perhaps because the value of the latter was
close to three trites (the virtual electrum stater), which might have made for more convenient accounting, although if so we have to ask why didn't Croesus go for this arrangement to begin with.
But in any case, and whatever the rationale for the weight of Croesus' staters may have been, it seems that we are now dealing with real value currency, or something close to it. Eventually this
process was completed when Darius reverted to the Persian weight standard for the daric and the siglos, presumably to establish the reputation of Persian coins as a full value currency at standard weights.
* The weight of the mina, and hence the shekel, varied quite a lot between different regions in the ancient world, and even within a defined region we can find a number of standards in use. The
weight of the Persian mina used here (c.500 gm) is taken from Tye's "Early World Coins & Early Weight Standards", p. 121; this value works well with the daric and revised siglos of the later 6th century. The
average weight of the daric as found is 8.33 gm, with a standard error of less than 0.01 gm; adding a small correction for wear, we can estimate the issued weight of the daric as 8.35-8.37 gm (assuming no
seigniorage, as seems to have been the Persian preference), consistent with a mina of 501-502 gm in the time of Darius.
The commonly used Euboic mina might also be considered as a source for Croesus' weight scale for the gold issues. This seems to have weighed c.430 gm in its Attic incarnation as a coining standard in the later 6th century (i.e, 6/7 of the Persian mina, as Herodotus reported, consistent with the weight of the Athenian tetradrachm of 17.2 = 430/25 gm, assuming no seigniorage). In Lydia the effective standard in Croesus' time (as judged from the electrum coinage) seems to have been somewhat less at 14.15 x 30 = 425 gm; this is equivalent to a stater of 425/60 = 7.1 gm, or perhaps 425/50 = 8.5 gm, but neither of these figures matches the actual weight of the light gold stater of 8.05 gm.
On the other hand, in his discussion of the Lydian coinage Mitchiner (in "Ancient Trade & Early Coinage") assumes a light "common Babylonian" mina of 409 gm and calculates a "shekel" weight of
1/50 mina = 8.18 gm, closer to but still distinctly heavier than the required 8.05 gm. Just where this particular mina value comes from is not really explained, but one place where similar values are found is
Phoenicia. There the weight systems, for precious metals at least, seem to have been based on the Egyptian/Phoenician beqa standard, which varied from 13 to 14 gm (as evidenced in the earliest double-shekel
coins of Tyre, for example). Depending on the city, the beqa then yielded equivalent mina (30 beqa) values ranging from c.390 to 420 gm, just the range that we are interested in. Theoretically, then, Croesus might
have adopted a similar mina standard and converted it to staters by dividing by 50, giving us finally the required value. However, tempting as this line of argument might seem at first sight, on the whole it is just
speculation, and I'm still not sure that we really know where Croesus' standard came from.
18 Aug. '13: Definitions of Types E and F revised.
14 Sep. '13: Type BE renamed "Evolving" (rather than Transitional) Type.
29 Dec. '13: Appendix on weight standards added.
15 June '14: Discussion of Massive Period revised.
23 July '14: Appendix on weight standards revised.
1 July '15: Possible silver 1/8th staters etc. noted.
9 Sep. '16: Discussion of heavy gold stater in Appendix revised.