There are problems with the reported statistics of the Latin Imitative trachy Type T. Leaving aside the Dolna Kabda hoard for the moment, there are only 9 large module versions of the type reported in
the various late Latin period hoards* (including a doubtful 6 from Krasen), as against 186 small module examples, a ratio of small to large types of 20:1. This is quite a large figure, but it is comparable to the
overall ratio for the Type A of Constantinople, the commonest coin of the later Latin period.
On the other hand, Jordanov's reported figures for the Dolna Kabda hoard give a very different picture. His figures are hard to untangle, but if I read them right, he reports 35 LM types out of a
total of 140 T's altogether, a ratio of small to large types of only 3:1, much lower than the figure above. However, when we look at the three illustrated "large module" coins in the hoard report, we see that
two of them (Figs. 42 and 46) are fairly clearly smallish module, with the design figures apparently only around 15 mm high, while the other (Fig. 44) may or may not be large module. Given that we know (from his
descriptions of the coins of John Comnenus-Ducas) that Jordanov tends to overate the nominal size of smaller module coins, we have to suspect that many of his 35 "large module" T's are in fact small to medium types,
and that therefore there the actual number of genuine large T's in Dolna Kabda is probably quite small, and possibly even zero.
The stray finds of Type T's reported by Jordanov in his booka on hoards and stray finds also show a low ratio of small to large types, but I'm prepared to disregard these as well, given that there are no convincing large module coins illustrated in the book (where the author is not loath to show clipped examples of large module types when nothing else is available). Much the same seems to be true of Dochev's (mostly stray) finds at Turnovo - Dochev reports 95 Type T's altogether in his bookb,
but doesn't break them down into large and small type (most are clearly SM), and the only example he shows is of a small type (although to be fair, any LM examples would probably have been clipped, and in this case
he may have preferred to show only a whole SM type**).
It seems clear, then, that the ratio of the SM to LM types for Type T is quite high, and in fact it is probably even higher than the figure we started with. If we restrict ourselves to hoards where
the assignments can probably be relied on, we find one LM type out of 53 altogether in the Peter and Paul hoard (actually only a part hoard), one out of 57 at Petrich, and none out of 13 in Bendall's "Thracian"
hoard of 1972, giving only two LM types out of 123 altogether, a ratio of small to large types of about 60:1.
In other apparently well reported hoards there was perhaps one LM type out of 6 T's at Serres, none out of 14 at Alexandrovo, none out of 16 at Nisovo and none out of 5 at Svishtov. There were also a
few small T's in other hoards, but no confirmed large versions. (Krasen reported 6 LM types out of 17 T's altogether, but there are other problems with size assessments in this hoard, so that these odd figures must
be suspect. Tri Voditsi/Dorkovo is not considered here as these are old finds which apparently did not distinguish large and small versions of the later types - "clipped" T's are mentioned but these may mean small
* For references to the various hoards discussed here see the article "Sequencing the Latin
LM Type T's from Bulgaria are invariably clipped - most of the published unclipped LM examples that we have were collected by Bertele in the western Balkans, where clipped coins generally did not circulate.
Putting all this together we see from the better evidence that we have a very high ratio of small to large types for the Type T of something like 50:1. This figure is extraordinarily high, and the
complete opposite of the situation for most of the other mid to later period types (i.e, Types D to R), where the small module versions are much less common than their large prototypes. Perhaps the figure is an
overestimate, but whatever the actual the ratio is, it is clear that the Type T was produced largely as a small type, primarily, we can assume, for use in Bulgaria. In fact, given that it is a relatively common
type, it is reasonable to suppose that the small T replaced the long running small Type A, which had been the mainstay of the Bulgarian currency for decades. (In stray finds of the Latin period in Bulgaria overall
the small Latin A makes up c.60% of all Latin coins, and c.40% of all coins, while the small Type T makes up c.4% of all Latin coins, and 2-3% of all coins).
However, we then have to ask where were the Type T's produced, since it is usually assumed that the later small A's were produced by the Venetians in a separate small module mint, very likely in
Constantinople (although some think it was located elsewhere). Perhaps by the time the T appeared, in the mid to later 1230's, the small module mint and the official Latin mint, both of which were probably run by
the Venetians anyway, had been recombined in Constantinople, due perhaps to the increasing impoverishment of the Latin rulers.
However, there is still another problem, in that the statistics for the Type T contrast sharply with those for the Type S, which is generally assumed to be a companion issue to the T. For the Type S
the overall figures are very much smaller, with only one SM version reported (at Svishtov) as against 5 LM types in other hoards. This ratio is the complete opposite of the figures for Type T, and is in fact similar
to ratios for the other mid to later period types. Evidently this type was not generally intended for Bulgaria, and hence it is probably reasonable to assume that it preceded the Type T, as an ordinary issue of the
official mint. (In stray finds there were 13 LM Type S's altogether reported by Jordanov and Dochev, allowing for possible double counting, as against no SM versions, but as we have seen the separation of LM and SM
types in these sources is suspect, so we shouldn't place too much reliance on these figures. In the D.O. and Ashmolean collections there are 6 LM and no SM versions).
For the other late types U, V and W the situation recalls that for the earliest Latin issues, with SM to LM ratios in the low single figures, so that the bulk of these types were evidently produced,
like the Type T, for use in Bulgaria. These types then could well have followed the Type T as issues of the central mint, although in what order is not clear. Similar considerations apply to the rare Jordanov Type
I, which seems to appear mostly in small module form.
On the other hand, we note that the rare late Hendy Types L and M are apparently known only in large module form; the significance of this is unclear, but perhaps it simply means that they postdate
1241 and the collapse of the Bulgarian empire.
a: I. Jordanov, "Coins and Coin Circulation in Bulgaria in the Middle Ages 1081-1261", Sofia 1984.
b: K. Dochev, "Coins and Coin Usage in Turnovo, XII - XIV C.", Turnovo, 1992.
29 June '08: Note on Types L and M added.
11 July '10: D.O. and Ashmolean figures for Type S noted..