The coins of Michael II of Epirus make up an interesting but little known subgroup of the later Byzantine coinage. They are all at best very rare, and some of the attributions are somewhat
questionable. Since there is little information available outside specialist literature on these types, this article seeks to summarise the current state of knowledge for the ordinary collector, and to assess the
reliability of the conventional attributions.
The Greek Despotate of Epirus.
In 1204 Constantinople fell to the crusaders. The "Despotate" of Epirus, as it later came to be called, was one of the remnants of the Byzantine Empire that remained under Greek rule after that
disaster, along with the "Empires" of Nicea and Trebizond. Epirus was ruled until 1215 by Michael I Angelus-Comnenus-Ducas, and then by his half-brother Theodore, who succeeded in extending his rule to
Macedonia, capturing Thessalonica from the Latins in 1224, thus turning the Despotate of Epirus into the Empire of Thessalonica. However, after he was defeated in 1230 by the Bulgarians under Ivan Asen, Theodore was
deposed and exiled, and Macedonia became a vassal state of Bulgaria, under the nominal rule of Theodore's brother Manuel. Epirus itself however, remained basically free, now under the rule of Michael II, son of
Michael I, and nephew of Manuel and Theodore.
Initially Michael was nominally subject to Manuel in Thessalonica, but from c.1236 he was generally acknowledged as ruler of Epirus in his own right, a position he maintained (with more than a few ups
and downs) until his death c.1270, although for much of the time he was essentially the vassal of John III in Nicea or Michael VIII in Constantinople. After Michael II's death, his eldest son Nicepheros ruled Epirus
proper as Despot, while an illegitimate son, John Doukas, ruled the region of Thessaly known to the crusaders as the "Duchy of Neopatras", which remained more or less independent until well into the 14th century.
The Coins of Michael II.
Despite his long rule, Michael II issued few coins, and it is now difficult to define them all with certainty, and almost impossible, in many cases, to determine their date of issue. So where do we
start? - for our purposes the most convenient treatment is that of Bendall, who summarised the standard views on Michael's coins in Num. Circ. 1996, pp. 3-5, listing one silver trachy and some 10 possible bronze
trachea (only seven of which are listed in Sear). Now some of the attributions of those types are uncertain, but others seem relatively clear cut, so let us begin with the latter.
(Note that it is generally assumed that coins of Epirus in the name of Michael are issues of Michael II rather than Michael I, except for Sear 2227, which is not considered here).
P.S. (July 2010) - Pictures of examples of many of the coins listed here can now be found in Eleni Lianta's excellent new book "Late Byzantine Coins 1204-1453 in the Ashmolean Museum University of
Oxford", Spink, London, 2009.
PP.S. More pictures and comprehensive line drawings are now also available in Marchev and Wachter's "Catalogue of the Late Byzantine Coins, 1081-1453, Vol. I", IVIS Ltd, Veliko Tarnovo, 2011.
As well images of examples of many of these types can be found in the online site of the Bibliotheque nationale de France (search on "Monnaie Michel Comnene-Doukas').
A: Silver ("electrum") trachy of Epirus.
Note that the ruler of interest on this and the following types usually seems to wear a short beard (at most).
Bendall 1 (S.2230, BnF BYZ-831).
Obv: Christ std on backless throne, r. arm raised in blessing, gospels in l.h.
To l, "IC/X/M"; to r, "XC/AK"
Rev: Michael stdg on l, and Constantine on r, in loroi, hldg lab. hdd sc. btwn; each also holds sc. cr.
To l, "(MI)XAH(Λ); to r, KW(NCTAN)TIN(OC)"
Here an attribution to a ruler named Michael is clearly justified, with the AK on the obverse possibly indicating Michael Angelus-Comnenus (rather than Michael Palaeologos), although another possible
interpretation might be Arta Kastra. Hendy's reading of the obverse sigla as IC(a)AK(oc), a supposed mint official, seems unlikely (to say the the least), seeing that it leaves the XC hanging. (In any case, the
exact interpretation of these sigla is not crucial to the attribution of this type. Similar sigla occur on the similar trachy S.2158 of Theodore C-D with a Christ obverse, but not on the other silver trachea of
Theodore with the Virgin on the obverse).
Note that this type is basically the same design as the "electrum" trachies S.2009/10 of Alexius III, but on a distinctly lower weight standard.
Note also that the dies for this type, and for the other issues of Epirus in finer style, notably Coins 2, 5 and 6 below, were quite possibly cut in Thessalonica for use in Arta.
B: Bronze trachea, probably of Epirus.
Note that most of the types in this section feature an AP monogram in the obverse field. It is tempting to think that this stands for Arta, but in some cases at least it occurs with a chi (X) and
clearly simply means archangel (as with the similar devices on the Thessalonican issues S.2183 of Manuel C-D and S.2165 of Theodore C-D); nonetheless it still helps to tie the various types of this period together.
Bendall No 2 (S.2231, Gr. 1278, BnF BYZ-832).
Obv: B. Christ, raising r.h. in front of chest. "IC-XC".
Rev: Ruler in lor. stdg on l, hldg sc. cr, and Arch. on r, both hldg sword btwn w. their left hands; Arch. also crowns ruler w. r. hand.
This type has been found at Arta (Arch. Delt. 24 1969, B', p.248, #21) and also at Aghios Achilleios in north-western Macedonia (ditto last reference). It was originally attributed to Manuel Comnenus-Ducas at Thessalonica (as Hendy's original Type E of Manuel), but on the Aghios Achilleios example the legend apparently reads as above, so that it would seem clear that this is an issue of Michael II at Arta. (Note that Bendall expands the legend to "O
Δ(ovkac)", although this is presumably not meant literally).
However, this type presents some problems. According to Jordanov there were no less than 7 examples of the type in the very late Latin period hoard of Dolna Kabda in north-east Bulgaria (Coins
1412-18), 6 of which were small module types, with one of medium module, and some of these were struck on cut down large module coins, apparently issues of Manuel C-D. If true, this would suggest that this type is a
coin of Manuel Comnenus-Ducas from Thessalonica, as was originally thought.
It is noticable, however, that these small types do not seem to appear in any of the numerous other later Latin period hoards, so there must be some doubt about the identifications of the Dolna Kabda
examples (which are all cut down to the very small Series III size)*.
For the present, then, Epirus has to be accepted as the source of this type, an idea which is supported by its rarity. It is also conceivable that this type was originally issued by Manuel in
Thessalonica, and then copied by Michael in Epirus under his own name in the late 1230's, but if so, where are the large module originals of Manuel? (which would, presumably, be more common than Michael's type).
Note incidentally that Grierson most inadvisedly later reused the Type E label for another quite similar issue of Manuel Comnenus-Ducas (S.2181), which can only lead to confusion - it certainly
confused me, and apparently Grierson as well, since on p.378 of his "Byzantine Coins" he labels coin 1208 as the (new) Type E of Manuel, when it is actually Type D (cf. p.262 for the correct descriptions).
* The pictures in the Dolna Kabda report are indecipherable. One of the Dolna Kabda coins appears somewhat more clearly in Jordanov's "Coins and Coin Usage in Medieval Bulgaria, 1081-1261"
(Sofia, 1984), as coin 11, Pl. XXXIII, but even there a positive identification is not possible. It is therefore difficult for us to make our own judgement on the coins in question, but it would seem unlikely
that Jordanov has misidentified all of them, and it is not clear what other type they could be.
Bendall 4. (S.2232, BnF BYZ-833 - reduced module type?).
Obv: B. Archangel, hldg jewelled sc. & orb.
In field to l, AP (ligatured); to r, M?
Rev: Michael stdg on l, in lor, hldg sc. cr. & aka, crowned by Virgin (not Christ) r.
The reverse legend is clear and confirms this rare type as an issue of Michael II.
The only published example of this type (the one shown in DOC IV and as No. 4 in Bendall's article) is broken, and seems a little small for a full sized trachy. From the data in Bertele's original
article on this coin (in Byzantinische Zeitschrift 1961) we see that it would have weighed c.1.9 gm originally, and had design borders 19mm across, figures which are on the lower limits for large module types. (The
supposed Ashmolean example Lianta 470 weighs 1.75g, but in fact this coin appears to be an example of No. 9 below).
Although the known example of this type is not particularly large, it is still noticable that tit is unclipped, which is one reason for attributing this type to Arta, rather than a more eastern mint.
Also, given that this coin came from the Bertele collection, it seems that it may have circulated in the western Balkans (rather than Bulgaria).
This type is very similar (but somewhat different in detail) to the scarce full size trachy S.2298 attributed to Michael VIII at Thessalonica (cf. CNG 66-1797), where the Archangel holds a trilobate
Bengal 5 (S.2235, Lianta 475, BnF BYZ-835, 840).
Obv: B. Archangel, hldg jewelled sc. & globus (without cross?)
In field, obscure symbols. (X/AP (lig.)-X/M, according to Hendy).
Michael stdg on l, wearing chlamys & domed crown (of despot?), hldg trilobate sc. (or palm leaf?) & aka?, crowned by John III Vatatzes (with forked beard), wearing loros & ordinary crown, and hldg sc. cr.
To l, "MIXAHΛ", To r, "IWENXW" (or subsets thereof).
The reverse legend is well established here, so that the characters are almost certainly Michael II and John III. There is no actual record of Michael being made Despot by John, but it reasonable to
assume that he was (as his son Nicepheros certainly was). If so, then possibly this type dates from 1252/3, when Michael was forced to submit to John III, after he had tried his luck attacking the expanding empire
of Nicea. (Hendy dates this type to 1248, supposedly on the occasion of the marriage of John's daughter to Nicepheros, although, according to Norwich, this marriage didn't actually take place until some eight years
later, i.e, after John's death).
Bendall 6 (S.2236).
Obv: Virg. h/l, orans. "MP-ΘV"
In field below l, A-P (Bendall), or X/Δ-X/P (Hendy).
Rev: As previous (exactly).
No examples of this type with a clear reverse legend are known, but otherwise the reverse is exactly as in the previous type, so it too is presumably an issue of Michael II.
Hendy, in a rather confusing passage in DOC IV, wants to make this type an issue of Thessalonica, on the basis of a supposed similarity of the sigla on the obverse with the D-P found on some of John III's issues (of Magnesia), but to others the sigla look more like A-P (perhaps meaning, in this case, Arta?). In any case both of the last two
types have apparently been found at Arta.
Hendy Pl. 41, 20-21, Sear 2225, Lianta 436-9.
Obv: B. Christ (beardless), with r. hand in front of chest; "IC-XC".
On some, "OE/MM/A - Nov/HΛ", or sim.
Rev: Arch. stdg in diamond pattern lor, hldg lab. (r. hand high) & gl. cr.
Blw, to l, X/AP (ligatured) & to r, X/M.
This scarce anonymous type (not included in Bendall's article) is very similar to the much commoner Latin Type P, but St Michael holds a full labarum and the reverse legends are at the bottom. Hendy
makes this type an issue of Thessalonica, on the basis of the fine style, and in DOC IV assigns it to Demetrius Comnenus-Ducas, on the basis, apparently, of who else? However, the type seems unlikely to be an issue
of this ruler at Thessalonica, given the fact that it is apparently only known in large module format (Hendy's typically dismissive comments on this objection notwithstanding - this type is generally heavier than
the coins of John C-D, and therefore could hardly, if it was an issue of Thessalonica, be later than John). In any case, if it is, as seems likely, the prototype of the Latin P, then it can hardly be later than the
mid 1220's, in which case it could well be an issue of Theodore C-D rather than Michael II.
According to Greek sources (e.g, I. Touratsoglu, Rev. numismatique, 2002, p.387), this type has been found at Arta, and it was also found at Aghios Achilleios in north-west Macedonia,
suggesting that it is an issue of Epirus (although perhaps the dies came from Thessalonica). As well, this type is not, as far as I know, found in any Latin period finds in Bulgaria or Thrace, and it also
seems that, unlike the other large module issues of Thessalonica, it is usually found unclipped, all of which supports a western origin for the type.
C: Bronze trachea, possibly of Epirus.
Bendall No 3 (S.2233, BnF BYZ-779).
Obv: B. Mil. St (Theodore) hldg large shield in each hand. There is also, it should be noted, a spear behind Theodore's left shoulder.
Rev: Emp. h/l in lor. and Mil. St (Demetrius?) h/l hldg long sword(?) btwn; manus dei abv. emp. from star-in-cloud sky abv. centre.
To l, ".....ΛΔEC", to r, unclear.
This curious type is very similar to the Thessalonican trachy S.2182 of Manuel Comnenus-Ducas, where Theodore holds a spear in his right hand (over his left shoulder), and a shield in his left hand.
The only thing that suggests this type is an issue of Epirus is its rarity - the only examples that I am aware of are the two coins shown in Bendall's article, and only one of these is convincing, namely the trimmed
large module example (the halved coin is possibly an S.2182). (The lightweight coin offered in the "Despot" sale is in my opinion probably an example of the relatively common small module version of S.2182, although
it sold for ChF 2800!).
The reverse legend reads ...ΛΔEC, as on S.2182, which suggests that this is a coin of Manuel rather than Michael (where the legend would normally read ..ΛOΔ(ovkac),
although it could still date from a time before c.1236 when Michael was possibly still calling himself Despot.
More likely though this type is just an odd variant of S.2182 from Thessalonica. The spear tip over the saint's left shoulder suggests that the die used was an S.2182 die recut with a shield in the
saint's right hand instead of a spear.
Ultimately, only an example of this type with a clear legend will solve the problem of attribution.
Bendall 11 (DOC IV p. 703, Uncertain No. 9, Type B of Arta?, BnF BYZ-845 - not in Sear).
Obv: B. Mil. St (Demetrius?) holding spear and scroll(?).
Legend obscure, but to r, A.
Rev: Two crowned figures w. short beards, both wearing loros, hldg patriarchal cr. btwn, w. the l.h. figure senior.
Stylistically, this large module type is perhaps closer to the routine issues of Thessalonica than to most of the types discussed so far. However, there was one example of this type in the well known
Arta 1923 hoard of the early 1260's (Class VIII), and other examples have apparently been found in the excavations at Arta. Given this, and the scarcity of this type, plus the "A" on the obverse in the same style as
on other Arta issues (e.g, Bendall 1 above), it is reasonable to conclude that this type is very likely an issue of Epirus.
However, given the lack of detailed legends, I will leave this type in the possible rather than probable category for the present. The identities of the characters are not known with certainty –
Mattingly suggested the type might a late joint issue of Michael II and Nicepheros, which at least fits the few facts we have to go on, while Hendy prefers Constantine for the second figure. The short beards
indicate that neither figure is John III.
Note that this type is generally very similar to S.2267 of Michael VIII with Constantine, but with differences in the details - e.g, here the left hand figure is senior, as we would expect if two actual rulers are shown. (Grierson's theory that the Arta 1923 coin was a badly struck example of S.2267 is clearly incorrect).
Bendall: Num. Circ. 2002, p.104, 2.
Obv: Christ std on backless throne; "?-XC".
Rev: Archangel stdg, hldg ? sc. in l. hand.
Another anonymous type (not in Bendall's 1996 article). Known only from a single clipped example, this type is included here because it is possibly, as Bendall suggests, a companion issue of S.2225
(see above), although the clipping suggests it could also be a Latin issue.
Another anonymous type which is possibly linked to these types is the rare (unique?) "Latin" Type Q. This is possibly an imitative type, as Hendy assumed, but its rarity and nonappearance in Latin
period hoards suggest that it may be an issue of Epirus (see "The Bulgarian & Latin Imitative
Types" for more details).
DOC IV p.702, Uncertain No. 8, Type A of Arta?
Obv: B. Christ; "IC-XC".
Rev: Ruler stdg on l, in chlamys, hldg jew'd sc? & ?, crowned by Virgin r.
Hendy suggests this type as a possible issue of Arta, although there is nothing except rarity to connect this type to Epirus, or anywhere else. It is very similar to one of the trachea* of Stefan
Radislav Dukas (1228-33) of Serbia, and it is not impossible that it is a Latin imitative type, based, as Stefan's type probably is, on coins of Manuel Comnenus-Ducas.
* Gr. 1279, a trachy in fine style, with B. Christ Emmanuel on the obverse, and czar in loros crowned by the Virgin on the reverse, with a full circular legend.
Bendall No 7 (S.2237, Lianta 476-7, BnF BYZ-841-3 - medium to large type mostly).
Obv: Castle (Walled City) with three towers.
Ruler stdg on l, wearing loros (sometimes diamond panelled) & ordinary crown, r. hand in front of chest (hldg nothing?) & aka. in l, crowned by ruler (or saint?) on r. wearing loros (sometimes diamond panelled) & ordinary crown, and hldg sc. cr?
Legend (on some examples):
To l, "MANovΛΔEC"; to r, "IWDEC"
It's worth noting that his type comes in two styles, one in normal fine Thessalonican style, and another in a cruder style. Whether these represent one issue or two separate issues is not
In Sear the reverse of this type is said to be the same as that of No's 5 and 6 above, but in fact it is different in detail and, as just noted, sometimes cruder in style.
The attribution of this type has always been, and still is, a problem. Although it is normally treated as a large module type, the type is also found in medium module (Series IIb) size, and there were
reportedly two very small module (Series III size) examples of the type in Penchev's Petrich hoard of late Latin period coins from southern Bulgaria (Coins 991-2)*. Taken at face value, all of this suggests
that this type could be an issue of Thessalonica.
According to Penchev (p. 90), one of the small Petrich coins showed the fragmentary legend "..ANI..", which if correct would confirm John ("Iwanic") as the left hand ruler, despite the fact that the
figure seems to show a short beard - otherwise he could be Demetrius. In either case it's then tempting to assume that the right hand figure is John III,
so that this type possibly belongs to the period 1242-44, when John C-D ruled in Thessalonica as despot under John III, or 1244-46 when Demetrius was despot.
However, there are problems with this idea. To start with, John C-D is usually shown with no beard, apparently because of his youth; however by 1242 he had probably reached puberty, which could
explain the beard on this type**.
Another problem with this type is the identity of the senior figure. It seems that none of the pictures of this type clearly show him with a forked beard (in fact they mostly clearly show him without
it), so there is definitely a problem making him John Vatatzes. On the other hand, we note that the right hand figure is named as IwDecpotic, which is a standard legend for Vatatzes (although also, in fact, for John
One final problem is the question of the mint. One the one hand we have small Series III size examples of this type, typical of the mint of Thessalonica under John C-D. But on the hand the more common
larger module examples are generally not found clipped, which would seem to rule out Thessalonica (or Bulgaria) as the source of this type, leaving only Arta, or at least a mint in the western Balkans, as the
source. The location of the mint for these types is, it seems, a problem yet to be solved.
But let us forget the problem of the mint for the moment, and press on with the problem of the issuers of the coins. If we rule out Vatatzes as the right hand figure, what other possibilities are
there? Hendy suggested one possible alternative. The close resemblance of the design of this type to the trachy S.2181 of Manuel C-D (with Constantine supposedly on the right) is obvious, suggesting that here we
might have a similar issue of John C-D. (This attribution might explain, incidently, the fact that this type is clearly commoner than the other issues attributed to Michael II, involving a reasonable number of
dies). Against this the general objection noted earlier to Thessalonica as the mint remains, and now of course we have the specific problem of the IwDec(potic) legend on the right, which Hendy was unaware of.
This leads us to another possibility. Some examples of this type show traces of a legend that could be read as "MANovHΛ(Δ?).." (cf. e.g, CNGe320-655, although on this coin the legend has
been partially recut with what seems to be a retrograde "MIX"). This suggests that this type is (or was initially) an issue of Manuel C-D (as overlord of Epirus at Arta in the 1220's, in which case the right hand
figure could be Manuel's ultimate overlord, Ivan Asen, who, unlike John Vatatzes, is usually shown with a short beard. As well, in the case of the particular CNG coin above, the recutting of the legend suggests that
the type may have been adapted to serve as an issue of Michael II and Ivan Asen at Arta after 1230 when Michael replaced Manuel as ruler of Epirus, as suggested by Popov#.
But here again there is a problem - ''IwDec(potic)'' is not a known (or likely) legend for Ivan Asen.
Overall, we seem to be left without a simple solution to the problem of the attribution of this type. Perhaps then there is a more complicated one - perhaps this design was used by two (or even more)
sets of rulers - perhaps the right hand figure is sometimes Constantine, sometimes Ivan Asen and sometimes even John Comnenus-Ducas. Different issues by different rulers (in different places?) might also account for
another feature of this type, the variety of sizes of the various examples that we have.
In any case the one thing that is clear is that with this type we have an attribution problem which is yet to be satisfactorily resolved, so that it might be more appropriate to include this type in
the "Uncertain" group below.
* The examples in the Ashmolean and DOC IV all seem to be of medium module, and a smallish module (1.7 gm) example of this type appeared in the Spinks sale in
N. Circ. May 1980 (Lot 2439). There is also a small module example in the Bibliotheque Nationale.
On the other hand I have seen five coins in the market weighing from 2.22+ gms (Coin 3319 in the Stacks 12/01/2009 Sale) to 3.0 gms (Goldberg 55-602), although the figures on some of these coins at
least are apparently only 15-16 mm high, typical of Series IIb, another of the exceptional features of this type. (Cf. Article "The Coins of John Comnenus-Ducas" for the Series IIb types in general).
See, e.g, Bendall and Protonotarios, N. Circ. 1978, p.178, Coin 2 (note). Note also that the Series III version of the small Latin Type A of Constantinople issued in the name of John C-D (Petrich hoard, coins 938-60) shows the emperor with a short beard.
There is also an example of S.2181 with a reverse legend which reads, according to Popov, "Tsar Asen" in some sort of Cyrillic script. This reading may be debatable, but the legend certainly isn't anything like Constantine.
D: Uncertain bronze trachea.
Coins 8 and 9 here were included in Bendall's list because examples of these types, and also of Coin 7, were part of a group of coins originally acquired by Bertele in the western Balkans which didn't
seem to fit anywhere else.
At first sight it might seem that the small types Coins 8 and 10 are probably issues of Thessalonica. However, there are difficulties with this idea. To begin with, they lack large module
counterparts, although this is certainly not a fatal objection. But furthermore, the ruler in all cases seems to have a beard, which might seem to rule out John C-D, while on the other hand the examples shown by
Bendall are not struck on cut-down large module types, as is often the case with the small issues of Theodore and Manuel C-D.
The medium sized Coin 9 is very similar to S.2232 and hence could well an issue of Michael II, but could also perhaps be an issue of John Comnenus-Ducas at Thessalonica, since, as we have seen
earlier, the fact that the ruler is bearded may not necessarily exclude John (or even Demetrius).
However, none of these types do not seem to have appeared in any late Latin period finds in Bulgaria or Greece, which again tends to militate against their being issues of Thessalonica. On the other
hand there is nothing, other than their apparent provenance, to specifically tie any of them to Epirus (or anywhere else), so that their attribution remains uncertain.
Bendall No. 8 (S.2234, Lianta 474, BnF BYZ-837-9 - small module type).
Obv: Star over crescent, "Π" (Pi?) abv?
Rev: Bearded? ruler stdg in loros hldg sc. cr. & aka.
Bendall thought that the small size of this type meant that it was a later issue of Epirus, but as just stated, there is no specific evidence to support this identification.
Bendall No. 9 (S.----, medium type?).
Obv: B. Arch, hldg jewelled sc? & orb?
Rev: Ruler 3/4 l. on l. hldg sc. cr. & aka, crowned by Arch? r.
Note that contrary to some opinions, this type is not the same as S.2332 above.
Bendall No. 10 (S.----, Lianta 471-3 - small module type).
Obv: B. Mil. St (Demetrius), hldg sword & sword hilt (not shield).
Rev: Ruler stdg in chlamys on l, hldg sc. cr. & aka, crowned by Christ r. Small star above and below crescent in right field blw.
To l, "X/M? EC"; to r, "IC-XC".
Another small type, like No. 8, to which it seems to be connected by the reverse symbols. This type is virtually the same as S.2184, the small module version of S.2177 of Manuel Comnenus-Ducas, but
the minor details are apparently slightly different. Thus on the "Michael" type the chlamys is jewelled, while on S.2184 it seems to be more plain (and certainly not diamond panelled as on the large module version
S.2177), and S.2184 as usually described supposedly lacks the added crescents in right field on the reverse.
The reverse legend of the "Michael" types is unclear but might (on one coin) read X/M (Δ)EC, in which case this could well be a late issue of Michael II. However, this reading is doubtful, and
doesn't really read like the other legends of Michael. Furthermore on some SM coins normally listed as S.2184, where the legend clearly reads Manouhl, we find the crescent between stars in the right field, as on
Coin 10 here, so this may in fact be a feature of S.2184 as well.
It therefore seems possible that the Bendall 10 coins are really just SM issues of Manuel Comnenus-Ducas (which is the opinion of C.L.B.C. 16.9), although it's conceivable the same type was issued by
both Manuel at Thessalonica and Michael at Arta.
The above coins clearly do not form a homogeneous group, and it seems that they may represent more than one ruler. Several of the coins can be confirmed as issues of Michael II of Epirus, but other
types could well be issues of Thessalonica, possibly by Manuel Comnenus-Ducas or his successors. The "Castle" type (Coin 7 above) is a particular problem - it seems more likely an issue of a western mint than of
Thessalonica, and the common attribution to John Vatatzes (and whoever) seems to be unsupportable.
The three uncertain types (Nos. 8-10) also present problems. Bendall 9 is very similar to S.2232 and hence is could be an issue of Michael II. Bendall 10 is probably just a version of S.2184 of Manuel
C-D (at Thessalonica presumably, although Arta is not out of the question), and Bendall 8 therefore is likely also an issue of Manuel.
(P.S. Some other possible issues of Arta, but probably of Theodore Comnenus-Ducas rather than Michael II, are discussed in the Note "The Arta 1983 Hoard").
20 Oct. '07: Bendall anonymous type added as possible coin of Epirus.
18 Feb. '08: Coin 2 discussion revised, again.
22 Feb. '08: DOC IV Uncertain Type 8 added as possible coin of Epirus.
15 Nov. '08: Coin 7 discussion revised, yet again.
19 Oct. '09: Coin 7 discussion revised, once more.
10 July '10: Ashmolean (Lianta) references added.
12 July '10: Coin 7 discussion revised, yet again.
20 Jan. '14: Ivan Asen as second figure on Coin 7?
21 Feb. '14: Coin 7 discussion revised, yet again.
14 Jan. '16: Coins 10 (and 8) reassigned to Manuel C-D.
11 July '16: Discussion of Bendall 4, 8, 9 and 10 revised.
28 Dec. '19: Discussion of Bendall 3 revised.