Lowick, N.M., S. Bendall, & P.D. Whitting. The Mardin Hoard. (London, 1977).
In 1972 the London firm A. H. Baldwin & Sons purchased a large collection of more than 13,000 uncleaned AE Byzantine folles from a source in Germany. The coins were mostly in quite worn condition but were nevertheless of particular interest because about 17% (that is, just over 2,200) carried Islamic countermarks. This was then (and remains today) by far the largest single known group of such Islamic-countermarked folles. And the original hoard from which these coins came was perhaps double the size of the assortment acquired by Baldwin 's, because another similar tranche had apparently been sold previously, split between a number of different buyers.
These coins came to be known as the 'Mardin Hoard ' (hereafter, MH). Mardin is a town in the south-east of Turkey near the Syrian border. In the 12th century Mardin was the capital of one principal branch of the Artuqid dynasty and several of the commonest countermarks represented in the MH can be ascribed to those Turkoman princes. It is quite plausible that the MH did originate in the area around Mardin. But 'Mardin ' is best understood as an educated guess, since no precise information about where the MH was found has ever been publicly revealed.
Before reselling the MH group, Baldwin 's arranged for a thorough numismatic analysis. The results were published as a monograph, The Mardin Hoard, in 1977.
Nearly all the MH coins were Anonymous Folles and other Imperial Byzantine issues dating from the period 969 – 1092, i.e. from the reign of John I Tzimisces until the currency reform under Alexius I Comnenus. The single commonest type was the Class K Follis of Alexius I (1819 specimens, of which 392 have countermarks); yet almost equally common was the much earlier Class A2/3 (1797 specimens, 170 countermarked).
There were a limited number of atypical types: about 130 of the coins predate the Anonymous Follis period and at the extreme there were a handful of 6th century folles (Anastasius, Justin, Justinian, etc.). A few contemporary counterfeits were also mixed in.
At the other extreme the MH included a single post-reform Alexius I tetarteron and a meager three Islamic coins. One could regard these as chance strays, but they are patinated and worn identically to other MH coins. The two latest Islamic types were Seljuk issues, dating from the early 13th century. Given the wear, this suggests that the MH may have been hidden during the Mongol disruptions of the second half of the 13th century. Also, taking into account the age and condition of the vast majority of hoard coins at the presumed time of deposit and the distinct minority of coins with countermarks, one might argue that the MH represented part of a copper bullion stockpile rather than a currency stash.
The countermarks are the most interesting aspect of the collection. The 1977 monograph also took into account some 500 examples of countermarked folles known from sources other than the MH and it remains the standard reference guide to Islamic countermarks on Byzantine coins. Nicholas Lowick identified 28 distinct countermark types, generally consisting of a few letters in Kufic or Naskhi script within a circular or rectangular border. Some types are represented by hundreds of examples, others by only a few – in one case by a single specimen. There are a significant number of coins bearing more than one countermark, and sometimes one countermark is over-stamped directly on another (which establishes relative dating).
The phenomenon of countermarking seems to have been limited in time to roughly the second half of the 12th century and limited geographically to the area including Mardin and
One relatively uncommon countermark (number 5 in Lowick 's assignment ) may be of special interest, because it consists of the Arabic word 'dhimam '. This should mean the right of a dhimmi, a non-Muslim, to enjoy the protection of the local Muslim ruler. Such protection was dependent on the payment of a special tax, the jizyah, which only non-Muslims paid. The authors of The Mardin Hoard speculate that the entire series of Mardin countermarks may have something to do with the collection of the jizyah, rather than with more general currency validation.
In 1997 a new hoard of about 2100 folles, of which 331 were countermarked, appeared on the London market, again quite worn and characterised by a hard green crystalline patina. These coins were studied and catalogued by Tony Goodwin.  Goodwin 's results largely confirmed the analysis presented by Lowick, Bendall and Whitting, which he extended by illustrating a series of variants, most notably in the writing of the word lillah (countermarks 16 and 17). Goodwin also noted two new countermark types, neither of which he could read or interpret with confidence; it is possible that these are blundered forms of types already present in Lowick 's 1977 list.
For a few examples, see my gallery: https://www.forumancientcoins.com/gallery/thumbnails.php?album=3405