Questions of Style and Detail in Byzantine Numismatics Metcalf, Michael Greek, Roman and Byzantine Studies; Fall 1960; 3, 4; ProQuest pg. 203 John Vatatzes and JohnComnenus Questions of Sty Ie and Detail in Byzantine Numismatics Michael Metcalf A HUNDRED AND TWENTY YEARS AGO there came to the hands of H. P. Borrell, a British merchant residing in Smyrna and a numismatist of distinction, a very large treasure of Byzantine gold coins that had been found in the region. The half-dozen pieces of Michael VIII which it included showed that the date of conceal­ ment of the treasure was shortly after 1260. All the rest of the nearly 1000 coins were identical in design. They were scyphate, or saucer­ shaped, and on the concave side showed two standing figures - to the left, the emperor, wearing a crown and loros and holding a labarum and akakia, and to the right, the Mother of God raising one hand in blessing as, with the other, she sets the crown on the emperor's head. The convex side of the coins showed Christ the King, his work accomplished, seated on a throne, with one hand raised in benediction and the other holding the Book of the Gospels. Half a dozen of these coins bore the name of Theodore II Lascaris (1254-58) while on all the rest on which the inscription could be seen1 it gave the name and titles, John Despot, the Porphyrogenete. Borrell, in publishing the discovery,2 questioned de Saulcy's attribution of these coins to John II Comnenus (1118-43). They belong in fact to Theodore's predecessor at Nicaea, John III Vatatzes (1222-54), as was indicated by the composition of the hoard, but lThey may be presumed to have been the great majority. The age-structure of the de­ posit suggests, as does that of the Smyadovo hoard mentioned below, that John's coinage was struck in fat greater quantities than that of Theodore. 2H.P. Borrell, "Unedited coins of the Lower Empire", Numismatic Chronicle 4 (1841- 42) ISH. 203 204 MICHAEL METCALF GRBS 3] Borrell's uncertainty was understandable, for the title Porphyrogeni­ tus was much used by John II whereas John Ill's claim to have been born in the purple was without substance.s The correct attri­ bution was given in the Sotheby sale-catalogue of Borrell's coin collection4 and also by Rollin in another article prompted by the hoard.5 Wroth, however, in the second part of the British Museum Catalogue of Byzantine coins listed the type, including some of Borrell's specimens,6 under John II. Three years later, in a com­ panion volume, he described them again as coins of John III, pro­ fessing himself finally convinced by the arguments for the later dating.7 It must seem to numismatists of the present generation that Wroth was slow to recognise the weight of the hoard-evidence, SIt was not until the end of his life that Theodore I named John, the husband of his daughter Irene, as his successor. 4Sotheby, 12 July 1852, lot 971. 5 [C.-L.J Rollin, "Monnoies d'or des empereurs de Nicee pendant I'occupation de Con­ stantinople par les princes croises, de 1204 a 1261," Revue Numismatique (1841) 171ff. Rollin, who was a much esteemed numismatic dealer in Paris, said that the coins he described had been found near Brusa, but it is tempting to suppose that Borrell's hoard and his were one and the same (cf. Bellinger's notes on the Corinth hoard of 1925 [infra n. 19]); the coincidence would be very great, seeing that no similar hoard has been recovered in a further 120 years. If this is correct, it would seem that Rollin's source may have had the first selection from the treasure; the exact quantities of each type must remain uncertain, as must the find-spot. Pace Mosser (op. cit. 79f infra n. 21) Borrell gives the date of discovery as 1839. 6W. Wroth, Catalogue of the Imperial Byzantine Coins in the British Museum, voL II (1908) 557ff. Nos. 32 and 34 are from the Borrell sale. It may be presumed, but is not certain, that they are from the Smyrna hoard. 7The same, Catalogue of the Coins of the Vandals, Ostrogoths and Lombards and of the Empires of Thessalonica, Nicaea and Trebizond in the British Museum (1911) 213ff. The catalogues are cited as BMC and distinguished from each other if necessary as BMCs and BMC3. The nomisma, Type 6, of Manuel I is also relisted in BMC3 where it is reattributed to Theodore I. Plate 9 - The Ibrahim Pasha Parcel No.1. Secret-mark, two crosses. 4.32g. BMCS 1. No.2. Very similar to No. 1. 3.73g. BMC 2. No.3. Very similar to Nos. 1-2. 4.00g. BMC 3. No.4. Two crosses (seven dots at shoulders). 4.50g. BMC 4. No.5. No mark. 4.55g. BMC 12. No.6. Diamond of four dots right. 4.26g. BMC 7. No.7. Die-duplicate of No.6. 4.63g. BMC 8. No.8. No mark (nimbus as Nos. 6-7). 4.52g. BMC 20. No.9. Dot left and right, close to seat of throne. 4.28g. (pierced). BMC 11. No.9 may well not have been associated with Nos. I-S. It seems quite likely that BMC2 John II (John III) 3, 4, 9 and 15-19 are from the same hoard. BMC 6 and 10, from the Ibrahim Pasha collection, are of John II, as is BMC 1, which is mounted for suspension. METCALF PLATE 9 1 2 5 8 THE IBRAHIM PASHA PARCEL PLATE 10 METCALF 10 11 12 13 THE ERYMANTHEIA HOARD [1960 JOHN VATATZES AND JOHN COMNENUS 205 but his reluctance over the title Porphyrogenitus was not without reason and points to further dimensions of the problem. Among the gold coins listed under John II in BMC there is considerable variety in style. Some of the specimens of BM C Type 3 - the type in question - are appreciably larger than the others, of finer workmanship, and with a more complete inscription in small, neat lettering.s Their style, which is quite easily recognisable, conveys an impression of space (PLATE 12, No. 30). It is found at its most characteristic only on the gold coinages of Alexius I Com­ nenus, 1081-1118 (PLATE 12, No. 28)9, and John II.10 Any doubts about the attribution of the large variety of BM C Type 3 are re­ moved by its occurrence in the Kastoria hoard,l1 which was concealed about the middle of the twelfth century. Type 3, then, is to be divided between John Comnenus and John Vatatzes/2 and we see that John III struck coins reproducing all the details of the design, even down to the legend In .6.E!IIO Tn IIOP<))TPO­ rENHT, of others issued by his predecessor of the same name a century earlier. This can hardly be coincidence. John III was, I believe, deliberately recalling the wise government and successful reign of John II13 and stating his claim, by the medium of a strictly imperial prerogative, gold coinage, to be the true inheritor of the BBMC 41-3; Pi. LXVIII, 11. 9BMC 4-8; PI. LXIV, 2, 3. l'BMC 1-2, 5-6 10-14, 41-3; PI. LXVI, 7, 8, 12, PI. LXVII, 1, 11. I1Bulletin de Correspondance Hellenique 71-72 (1947-48) 393. HBMC 41-3 are of John II, and 20-40 (= 1-21 among BMC3 1-2't) are of John III. 13An exact comparison is afforded by the young Eadgar, first king of England (959-75), who, very conscious that he was the great-grandson of Alfred, recalled the glories of the past by reproducing the design of his predecessor's pennies with the monogram of London. One wonders in both instances how the monarchs obtained the old coins. See R.H.M. Dolley and D. M. Metcalf, "The reform of the English coinage under Eadgar", in Anglo­ Saxon Coins (ed. Dolley), 1961. Plate 10 - The Erymantheia Hoard No. 10. No mark. 4.24g. No. 11. No mark. 4.37g. No. 12. Dot right (radiate nimbus cruciger). 4.61g. No. 13. No mark. Late style, e.g. throne? 4.30g. No. 14. Uncertain letter right. Late style? Of a more coppery ap­ pearance than the rest of the coins in the hoard .. 4.47g. No. 15. Dot .left, two dots right. Late style? 4.67g. No. 16. DIamond of four dots nght. Poor style. 4.46g. No. 17. Asterisk right, perhaps also an uncertain mark left. 4.24g. No. 18. Dot left, four faint small dots right. 4.40g. 206 MICHAEL METCALF GRBS 3] Empire, whose metropolis and western themes he saw it as his destiny to recover from the Franks and from the pretending dyn­ asties established at Salonica and in Epirus. His rptKEcpaAa, so called because of the three heads shown on them,14 were evidently issued in great quantities. They had a wide circulation in the coastlands of the Aegean and in the Black Sea hinterland of eastern Bulgaria, as is indicated by the various hoards in which they have been found. A large deposit was discovered in 1940 at Pirgovo on the lower Danube15 and a dozen more coins at Nesebir (Mesembria) in 1933.16 Twelve were found in a hoard at Smyadovo in northeastern Bulgaria in 1945 along with gold of Michael VIII and of Andronicus II with Michael IX (1295-1320).17 Bansko .i i. 'l. Salonica 1'. Krestiltsi Drama i. o ... Pidroasele i. Stoene~ti ·Pirgovo LAC K SEA MAP SHOWING THE FINDSPOTS OF GOLD COINAGE OF JOHN III The attribution of the coins in the Messini, Athens, and Stoenesti hoards is unconfirmed. The distribution of finds suggests the importance of two valley routes: (i) northwestwards from Nesebir, (ii) northwards from Salonica. 14The name was sometimes written in the abbreviated form rICh': see F. Dolger, "Chronologisches und Prosopographisches zur byzantinischen Geschichte des 13. Jahrhun­ dert," Byzantinische ZeitschriJt 27 (1927) 291fI. l5T. Gerasimov, in IztJestiya na Buigarskiya Arkheologicheski lnstitut' (hereafter IBAI) 14 (1940-42) 282ff, where the coins are attributed to John II. Dr. Gerasimov has kindly informed me that they are of the variety BMC2 20-40, and belong in fact to John III. l6The same,IBAI 8 (1934) 467ff, and the same remark. l7The same, in IztJestiya na Arkheologicheskiya Institut (hereafter IAI) 17 (1950) 316ft, and cf. lBAI 15 (1946) 235ft. [1960 JOHN VATATZES AND JOHN COMNENUS 207 From Greece, coins of the type from the three hoards of Eryman­ theia 1955, Drama 1949 and Thessaly 1949 are discussed and illU3- trated here. IS Two TpLKE4>aAa of John with one of Theodore II were found in the course of the Corinth excavations of 1925,19 and another came to light at the same place in 1934 in a hoard of French silver coins deposited after 1253.20 A hoard of 100 coins was discovered at Pergamum in 1912.21 Certain other finds, from which coins with the name of John were attributed to other rulers, may perhaps also have belonged to John Vatatzes.22 Wroth, in his later comments on BM C Type 3, remarked that there were other coins apparently of John II which ought perhaps to be transferred to John III.23 A number of them are of BMC2 Type 2, the commonest of the issues of gold by John Comnenus, which shows half-length figures of the emperor and the Mother of God as its principal design. The type, like Type 3, divides readily into two varieties, a larger, in the fine, early twelfth-century style re­ ferred to above (PLATE 12, No. 29), and a smaller, of inferior execu­ tion (PLATE 12, No. 31). Some if not all of the smaller pieces are obvious candidates for a thirteenth-century dating. The stumbling- 18The accession of the Erymantheia hoard is noted in BCH 80 (1956) 228; the 9 coins now in the Greek National Numismatic Collection were selected from 17 coins that were found. A similar hoard was said to have been found in the Patras district in the 1930's. The Drama hoard is noted in BCH 74 (1950) 292f. The gold coin from Thessaly wa~ found along with 46 VenetIan grossi and 7 fragments in a deposit concealed ca. 1256-60. Mrs. E. Varoukha-Khristodhoulopoulou, the Keeper of Coins, most kindly gave me every facility to study all three finds. Work in Athens was made possible by the tenure of the School Sudentship of the British School of Archaeology at Athens. 19A.R. Bellinger, Catalogue of Coins found at Corinth, 1925 (New Haven 1930) 74. 20See K.M. Edwards, "Report on the coins found in the excavations at Corinth during the years 1930-35," Hesperia 6 (1937) 241ft 21Note by K. Regling, in S. MeA. Mosser, A Bibliography of Byzantine Coin Hoards (New York 1934) 65. 22Messini, 1900, 4 coins (Mosser, sub Ithome); Athens, 1928, 10 coins (ibid.); Princes Island, 1930, 16,000 coins (Rassegna Numismatica 27 [1930] 150); see Mosser, Ope cit. 70 under "Prinkipo," but may there be a confusion with a hoard not of Byzantine coins? Cf. K. Regling, Der griechische Goldschatz von Prinkipo, Museum der Altertiimer zu Istanbul (Berlin 1931). The Stoene~ti hoard, described as of John II, is perhaps of John III; see B. Mitrea, in Dacia N.S. 2 (1958), 493ff. For the Krestiltsi hoard of 1952, 10 coins of John III, Type 3, and for the Preslav grave-find of 1953, 2 similar coins, see T. Gerasimov in IAI 20 (1955) 602ff. For the Bansko hoard of 1957, of 8 similar coins, see Gerasimov in IAI 22 (1959) 151. For a grave-find of one (pierced) coin at Pietroasele, see O. Iliescu in Studii fi Cercetari de Numismatica 2 (1958) 455. 23BMC' 3, 4, 7, 8, 9, 15-19. See BMCs, p. 215. 208 MICHAEL METCALF GRBS 3] block is that there are no provenances by which the matter could be settled. Among all the hoards of John III, Type 2 has never been recorded in association with Type 3. It has been said that Asia Minor is numismatically more like a whole continent than a single coun­ try.24 This is almost certainly true of the Byzantine coinages of the second half of the twelfth and the first half of the thirtenth cen­ turies; it seems to have been the place of origin of many of the scarce types of scyphate bronze coinage, for which provenances are nowhere else noted.25 The lack of hoard-evidence for Type 2 should not, therefore, seem an insuperable obstacle, although it indicates that the issue, if it in fact belongs partly to John Vatatzes, must have had a restricted currency. Even if only a few specimens could be shown to belong to John III, it would be enough to establish that he followed the remarkable monetary policy of reproducing not one but two of the designs of John II's coinages, and to make it necessary to consider carefully any coin apparently of John II in a style that would be exceptional for the twelfth century. The two specimens for which the best case can be made out are BMC 8 and 9 (PLATE 12, Nos. 32, 33). They are of small size, and their general style is unlike that of the best coins of John II. They have the bevelled edges characteristic of Nicaean issues. Both have secret-marks in the spaces above the throne on the convex side, a feature common on the gold of John III but unknown on that of John II. The throne itself, with jewelled sides, is exactly like those on the thirteenth-century gold, and quite different from the twelfth-century version, in which care is always taken to show Christ's footstool (PLATE 12, Nos. 28-30). The beginning of the inscription, KE BOH®EI, distinguishes BMC 8 and 9 from coins of Type 2 in twelfth-century style, on which it is lacking. The variation in weight between the two coins suggests a degree of carelessness in their manufacture which would not have been 24M. Grant, The President's Address, Tournai of the Royal Numismatic Society (1956) p. 4 (bound with Numismatic Chronicle, Sixth Series vol. 16 [1956]). 2liThe evidence from which I believe that this conclusion can be drawn is set out in "Byzantine scyphate bronze coinage in Greece," Annual of the British School of Archaeology at Athens 55 (in press). [1960 JOHN VATATZES AND JOHN COMNENUS 209 tolerated under John II.26 Finally, a detail in the ornamentation of the emperor's crown speaks strongly for a thirteenth-century dating. The pendants at either side are shown by vertical lines ending in a group of dots representing jewels. The form of the pendants is an important criterion in all the Byzantine coinages of the twelth and thirteenth centuries. Under the Comnenian emperors they are shown on the coinage in the best styles, almost always, by three dots arranged thus: '.' The standard form was changed towards the end of the twelfth century, to become : on the gold of Alexius III. The latter arrangement is common on the gold of Nicaea but after 1204 the minutely careful regulation of the coinage was not main­ tained. Half a dozen forms of the pendant can be found on John's TPLK€cPa'Aa.27 The dilemma of BMC 8 and 9, which have pendants of only two dots thus : , is that either they are coins of John II with so many exceptional features of style that they cannot be thought to have been struck at the metropolitan mint, or they belong to John III and are deliberate copies of the type issued a century be­ fore. There need, I believe, be no doubt which is the correct answer. The form of the pendants is sufficient to show that BMC 15-19 should also be transfered to John III: among five coins, there are as many different arrangements of the jewels. Type 1 is less readily divided on grounds of general style. BMC 3, 4 and 7, however, have pendants which make it certain that they are not coins of John II. BMC 1, 2, 5 and 6 would seem to be cor­ rectly attributed to the twelfth century. The coins to which Wroth drew attention should thus all be transferred to John III. It only remains to point out that the coin inscribed ®EOLln.PO~ LlE~II 26John III's gold fully maintains the traditional weight of the nomisma (although with only 2/3 gold content - see Rollin, op. cit., for assay results) but there was distinctly greater variation in weight from coin to coin. Means and standard deviations: (i) for Alexius I: the Depentzikos hoard, 9 coins, 4.34g., 0.097g. (ii) for John II: BMC 10-14 and 42, 4.34 g., O.088g. (iii) for Manuel I: the Corinth hoard, 4.32g., O.120g. (iv) for John III: my Nos. 1-8, 4.32g., 0.29g.; the Erymantheia hoard; 4.42g., O.l4g.; BMC2 8, 9, 15-19, 4.33g., O.24g. 27 A fuller discussion of the representation of the pendants will be found in the article on scyphate bronze coinage cited above. Under Manuel I, pendants of only one or two dots are by no means uncommon on the gold coinage; see the plate accompanying J.M. Harris, "A gold hoard of Corinth," AlA 43 (1939) 268ff. On gold of John II I have never seen any forms of pendant other than the 3 drop-shaped jewels on "stalks" as shown on PLATE 12, Nos. 29-30, 3 simple dots, or something transitional between the two forms (e.g. BMC 12 and 13). 210 MICHAEL METCALF GRBS 3] Tn ITTPOrE which Wroth gave to Theodore r28 is more plau­ sibly to be interpreted as a transitional issue struck at the beginning of the reign of Theodore II. There is a considerable range of style, although nothing like the contrast that has been described above, among John Ill's coinso Also, every other specimen has a secret-mark, indicative no doubt of some detail of the mint-organization behind its issue, above the throne on the convex side.29 At least a dozert such secret-marks occur. Their variety suggests that the organization of the coinage was both careful and complex, while the differences in style between one coin and another indicate either that there was a long stylistic progression during John's thirty-twa-year reign or that there was more than one mint at work - or both. Excluding one or two pieces which stand aside from the rest as being evidently provincial, such as PLATE 10, No. 17 for which the dies were engraved by a work­ man who did not understand how the loros was worn, there are: one or two small clues, such as the blundering of the legend Tn ITpr to become Tn IIPT, Tn IIPTPOr, etc., and secret-marks composed of initials such as ~P, which point to the existence of a number of mints. The problem is as intriguing as it is difficult; and the assertion that the precious metals were coined at provincial mints must still seem unorthodox enough to call for thorough proof.3o Only hoards could provide conclusive evidence by associ­ ating certain styles with particular regions. Until several hoards of John's gold coin3ge discovered in western Asia Minor have been described there will be no possibility of writing a definitive account 28RMC3 , p. 204f. 29The use of secret- or privy-marks became widespread on medieval coinages from the second half of the thirteenth century. Cf. the Venetian grossi of R. Zeno (1252-68) as listed by A. N. Papadopoli, Le Monete di Venezia descritte e illustrate (1893-1919) ad lac. Secret-marks to which those on the gold of Nicaea are perhaps similar in character were already in use on the byzantine bronze coinage under John Zimisces and his successor~ in the early eleventh century (Anonymous Type A); I have argued that they are localized in their occurrence and must therefore be mint-marks: see "Provincial issues among the Byzantine bronze coinage of the eleventh century," Hamburger Reitrage zur Numismauk 5 (in press). 30lt has become clear that not all the bronze coinage of the twelfth century was the work of the metropolitan mint. For the proposed attribution of certain varieties to Greek mints, see "Byzantine scyphate bronze coinage in Greece," loco cit. Wroth, BMC3 p. lxxii, suggests Nicaea as the mint-place of John III's gold. [1960 JOHN VATATZES AND JOHN COMNENUS 211 of its issue. At the moment provenances can be attached to very few specimens of which photographs are available.3! Part of a hoard which I believe must have come from Asia Minor is now preserved, although without record of provenance, in the British Museum. In 1849, the British Consul-General in Egypt, Sir Charles Murray, negotiated the purchase for the Museum of a considerable number of classical, Byzantine and Arabic coins from the collection which had been formed by the late regent, Ibrahim Pasha.32 Among them were nine coins of John III, at least eight of which have in common certain trivial points of style that distin­ guish them from the great majority of other specimens of the type and make it certain that they are from a single source. The points are so uninteresting in themselves that the coins would not have been selected from a larger number because of them. A comparative study of these trivialities, however, quickly shows that the variations, at first sight meaningless, are amenable to order. On all nine coins, the design on both sides is enclosed by a double circle of dots. ~10st other specimens are less carefully made 2nd have a linear border (compare PLATE 9, the Ibrahim Pasha parcel, with PLATE 10, the Erymantheia hoard), which may be double, but is more usually single and incomplete. A second char­ acteristic detail of the Ibrahim Pasha group of coins is in the row of 3 dots on the emperor's shoulders: on other varieties there are usually at least half a dozen dots (if the reader will continue to refer to PLATES 9 and 10, the coins will gradually begin to look less identical!). The 3 dots at the shoulders are a more general criterion than the secret-marks, for they are found with 2 crosses (PLATE 9, Nos. 1-3), a diamond of 4 dots (Nos. 6-7), and on coins without secret-marks (Nos. 5, 8). They are not, however, an invariable feature. This is shown by PLATE 9, No.4, which is closely similar to PLATE 9, Nos. 1-3 in such details as the loros hut has 7 dots at the shoulders. No. 4 must have come from the same workshop as 31The only hoard which has ever been properly published, with photographs, is that from Corinth in 1925. 32The transaction is recorded in a copy of a letter from Murray, dated 17 November 1849, in the manuscript Minutes of the Department of Coins and Medals, vol. I (1838- 56) p. 158, kept in the British Museum. Ibrahim Pasha (1789-1848) was commander­ in-chid of the Turkish forces at the Battle of Navarino. 212 MICHAEL METCALF GRBS 3] Nos. 1-3 but is perhaps a little earlier or later in date. Nos. 6-8 can be grouped together because they have five dots arranged in :it. cross, instead of only one, in each arm of the nimbus cruciger. The two with a secret-mark are struck from the same pair of dies; the presence of die-duplicates as well as the "near-duplicates" Nos. 1-3 in a hoard drawn from so extensive and varied an issue as John Ill's coinage suggests that those particular coins had not long left the mint. It follows that the Ibrahim Pasha hoard is unlikely to have been an AuslandsfundH from Bulgaria or Frankish Greece. The Erymantheia hoard is far more varied in character, and includes half a dozen different secret-marks. Nearly all the coins belong to the general group with a single linear border. Apart from the specimen with an asterisk as secret-mark, which stands quite alone (PLATE 10, No. 17), the only obvious intruder is a coin of neat workmanship with a double linear border (PLATE 10, No. 18). In order to find parallels to the varieties in this and the Drama and Thessaly finds one must, until other hoards have been published, turn to museum holdings and sale-catalogues. Three comparisons will be briefly made. The coin found in Thessaly in 1949 and illustrated on PLATE 11, No. 21 has badly blundered pendants and combines with a double linear border the rather unusual feature of a double linear outline to the cross in the nimbus. Other specimens with the same form of nimbus cruciger match the Thessaly coin in the double linear border and the blundered pendants which may even not be a pair. 33A find from territory which was outside the jurisdiction of the authority issuing the coins which it contained. There does not seem to be a convenient word in English. Plate 11 - The Drama and Thessaly Finds, and Coins Illustrating Stylistic Groups No. 19. No mark. 4.52g. Drama, 1949. No. 20. Dot in field left. 4.25g. Drama, 1949. No. 21. II right. The metal rather whiter than usual. 4.49g. Thessaly, 1949. No. 22. Uncertain, but perhaps pyramid of three large dots left (double linear outline to cross in nimbus). Foreign Ambassador sale, 704. No. 23. Dot left and right (radiate nimbus cruciger). 4.47g. BMC3 10. No. 24. Dot right (radiate nimbus cruciger). Foreign Ambassador sale, 690. No. 25. The same. H. P. Hall sale, London, 16 November 1950, 2306. 4.53g. No. 26. Uncertain mark right. Foreign Ambassador sale, 691. No. 27. The same mark. 4.25g. BMC 19. METCALF PLATE· 11 19 20 26 THE DRAMA AND THESSALY FINDS, AND COINS ILLUSTRATING STYLISTIC GROUPS PLATE 12 METCALF 29 33 ALEXIUS I, JOHN II; OrnER COINS OF JOHN III [1960 JOHN VATATZES AND JOHN COMNENUS 213 The Thessaly coin has the secret-mark II, and another from the Foreign Ambassador sale, unfortunately in very worn condition, seems to have three large dots arranged in a pyramid as a secret­ mark to the left of the throne (PLATE 11, No. 22)3\ while BMC 18 is unmarked ;35 as with the Ibrahim Pasha parcel, a single stylistic group includes more than one secret-mark. Second, a variety in the form of the nimbus cruciger in which the arms of the cross are exaggeratedly radiate points to another small group of coins which correspond with one another also in having pendants in the form : (although one of the dots is some­ times missing) and the secret-marks of a single small dot to the right of the figure of Christ or alternatively two dots, one at each side.36 The loros is of a variety which is practically standard, down to the last dot, among coins with a single linear border. A coin from the Erymantheia hoard belongs to this group (PLATE 10, No. 12) as do BMC 10 (PLATE 11, No. 23, two dots,) BMC 24 (one dot) and BM C 23 (no secret-mark). Two similar coins were sold in Munich in 195737 which, it is tempting to suppose, may have derived from a single hoard. Third, another coin from the Foreign Ambassador sale38 takes on added interest because its puzzling secret-mark, more easily recognised than described (see PLATE 11, No. 26), can be matched 34Glendining, 7 Match 1957, lot 704, incorrecdy catalogued as Theodore II. For the secret-mark, which is conjectural, d. John II (in fact John III) BMCs 8. 35It has two dots in the middle of the staff of the labarum and, unusually, two at the bottom. 36Cj. No. 9 in the Ibrahim Pasha parcel, where the two dots are much nearer the seat of the throne. 37K. Kress, 6 December 1957 (catalogue no. 106) nos. 522 and 523, listed as from different sources; incorrecdy described as coins of "Theodosus III" (sic). 38Lot 691. Plate 12 - Alexius I, John II; Other Coins of John III No. 28. Alexius I, BMC Type 1. Foreign Ambassador sale, 688. No. 29. John II, BMC Type 2. Foreign Ambassador sale, 696. No. 30. John II, BMC Type 3. BMC 42. No. 31. John III, "Type 2". Foreign Ambassador sale, 695. No. 32. John III, "Type 2", variety with thrones as on Nos. 1-27. Pyramid of three dots right. 4.8Ig. BMC 8. No. 33. The same variety as No. 32. Asterisk left and right. 4.l0g. BMO 9. 214 MICHAEL METCALF GRBS 3] nactly from BMC 19 (PLATE 11, No. TJ) which agrees also in the double linear border and the pendants. The way of working at a problem such as John's gold presents is very much like solving a jig-saw puzzle: here are two or three pieces that can be interlocked, here are several dozen that must somehow belong together in the same part of the finished picture. It is chiefly persistence that is called for, and the details of solving are of general interest only as an illustration of numismatic meth­ od. The point of recording them is that the pieces are not supplied but have to be collected, and that they are at present being neglected or even thrown away as hoards are summarily published or dispersed. What sort of picture will finally be put together it is too early to say, but the fragments which have been discussed will, I hope, con­ vince others who may have the opportunity to study John's gold that there is a picture to complete and that progress is not difficult. ST. JOHN'S CoLLEGE, CAMDIDQB September 1960