Analysis and Provenience of Minoan and Mycenaean Amber, III. Kakovatos Beck, Curt W;Fellows, Constance A;Adams, Audrey B Greek, Roman and Byzantine Studies; Spring 1970; 11, 1; ProQuest pg. 5 Analysis and Provenience of Minoan and Mycenaean Amber, III. Kakovatos Curt W. Beck, Constance A. Fellows, Audrey B. Adams I N 1907 Dorpfeld discovered three tholoi at Kakovatos1 which he equated with Homeric Pylos. This identification has long since been abandoned, but the site retains considerable interest, not least because the amber "spacers" found there invite comparison to those from Germany and England. The first preliminary report2 mentions finds only from the largest tholos, later designated as Tholos A, among them "amber which in quantity and variety exceeds anything known so far from Mycenaean tombs. In addition to lentoid beads of various sizes-the largest will have served as sword pommels-there are flat rings with broad loops for suspension, as well as rectangular plates with multiple drillings." The tomb was tentatively considered H contemporary with the older tholoi [at Mycenae] which immediately follow the shaft graves [also at Mycenae J." In the full report Dorpfeld3 concluded that the large Tholos A was the most recent, Tholos B slightly older, and Tholos C the oldest, but that all fell into the period of the "second late Minoan style," to which vases in all three tholoi belong. He based his assignment of relative age in part upon the fact that finds of amber and ivory are much more plentiful in Tholos A, while glass and stone predominate in the other two. The tholoi are now assigned to LH II A, or 1500-1450 B.C.4 The amber finds from Kakovatos were finally described, though 1 British War Office map series C.S.C.S. 4439/MDR 610 (1 : 100,000) sq uare L.5 (Kiparissia). grid lines 514E/998N. Cf R. Hope Simpson, A Gazetteer and Atlas of Mycenaean Sites (London 1965) no. 255. I W. Dorpfeld (with description of finds by K. Miiller), AthMitt 32 (1907) vi-xvi. 3 Dorpfeld, AthMitt 33 (1908) 294-317. 'A. Furumark, The Chronology of the MycetUlean Pottery (Stockholm 1941). 5 6 MINOAN AND MYCENAEAN AMBER, ill. KAKOVATOS still rather summarily, by Muller» in 1909. We now present, with minimal annotations, his description of the amber finds, followed by an account of the amber we were able to see at the National Museum in Athens in the summer of 1966 and of the results of our analyses. Tholos A: Beads Muller reports:6 "Tholos A was especially rich in amber, both with regard to quantity and to shape. The greater part is comprised of beads. In spite of the vulnerability of the material, about 500 beads were preserved entirely or in large part. Shape and size vary considerably as shown in the illustration [Miiller shows 10 beads on his Plate xv, nos. 10-18, 25, in natural size. Their diameters range from 5mm to 41mm]. The majority are lentoid and have sharp edges. Their con­ vex surfaces vary in curvature. There also are-but only among the smaller beads-flat plates with rectangular cross­ section. More frequent are medium-sized pieces of this kind with rounded edges. When beads of this kind are very thick relative to their diameters, they occasionally approach spheri­ cal shape but never quite attain it. The diameter of the drill holes also varies, both absolutely (in one case 8mm, usually between 5-!- and 1tmm) and relatively; in some smaller beads the ratio [of drill hole to diameter of the bead] is such that they appear as rings. "The dimensions vary even more. The diameter increases continuously from 5 to 50mm, and from fragments one can infer center pieces of which the largest has a diameter of 85mm and a thickness of 21mm. [Depending on the curva­ ture, a bead of these dimensions would have weighed between 40 and 80g.] "No beads of such size have been found so far in Mycenaean contexts. None of the beads from the shaft graves [of Mycenae] measures more than 40mm. It is difficult to imagine such giant beads strung into necklaces, but there is 6 K. Miiller, AthMitt 34 (1909) 269-328. • Miiller, op.dt. (supra n.5) 278-79. CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 7 no alternative. One might think of pommels for daggers and swords [as Miiller himself had done in 1907], but apart from the fact that they form a continuous series with the smaller beads, this is ruled out by the small drill holes-that of the largest piece measures barely 3mm-and the consideration that known pommels are never lentiform and were attached to the blade in a different manner." All the amber objects from Tholos A are registered in the National Museum of Athens under inventory no. 5688. A large number of beads and some of the more complex objects and fragments have been mounted on a board for exhibition in the Mycenaean Room. The beads have been strung into two necklaces; the first consists of 115 graduated beads of fairly large size ranging from 14mm diameter for the smallest to 42.5mm diameter and 20mm thickness for the largest center bead. Most of the beads are lentoid (in Beck's7 classi­ fication: Short Convex Bicone, group B.l.e.), though in many the curvature is so slight that they approach a bicone (Beck's Short Bicone, group B.2.e). A few of the smaller beads are discs, i.e., they have parallel flat sides. Their edges may have been sharper once than they are now and the distinction between Miiller's "flat plates with rectangular cross-section" (Beck's Cylinder Discs, group A.2.b or Short Cylinders, group B.2.b) and those with "rounded edges" (Beck's Barrel Discs, group A.l.b or Short Barrels, group B.1.b) may reflect differ­ ences in wear and weathering rather than in intentional, original shape. The second necklace on the exhibition board consists of 291 smaller beads ranging in diameter from 5.5mm to 18.5mm. Again the larger beads are biconvex or biconical and the smaller ones, almost half of the total number, are discs with sharp or rounded edges. In a cardboard box in the storeroom at the National Museum there are several hundred fragments, many of which are large enough to be recognizable as having been part of beads, pendants or spacers. To preserve the whole beads, samples were taken from eleven of the fragmentary beads. Their spectra, over the range from 7.5 to 9.5 fL, are shown in Figure 1 and the results of their analyses are listed in Table 1. By computer classification, ten of the eleven spectra were identified as those of Baltic amber; the exception, spectrum 1553 of sample 11, differs only minutely from the others but fails to meet the stringent computer test. 7 H. C. Beck, Archaeologia 77 (1927) 1-74. 8 MINOAN AND MYCENAEAN AMBER, III. KAKOVATOS 1543 1546 1549 ISS2 1544 8 ·9 8 9 8 9 8 9 ,.. Figure 1. PARTIAL INFRARED SPECTRA OF AMBER BEADS. KAKOVATOS, THOLOS A Tholos A: Pendants Muller continues:8 "In addition to the beads, the tomb contained the remains of a particularly splendid necklace. Its members have the form of flat round plates with a very broad suspension tube along one side and a circular hole which is so placed that the ring has its greatest width opposite the tube. It [the ring] is tapered on both sides towards the periphery, terminating into a sharp edge. Three pieces have been preserved entire; there are fragments of 5-7 others. The largest, which will have been 8 MUller. op.cit. (supra 0.5) 279-80. CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 9 the center piece, is 64mm high and nearly as wide with a thickness of 12mm; the smallest is 30mm high. The width of the ring increases to a maximum of 21mm in the former, and to lOmm in the latter." TABLE 1. SPECTROSCOPIC CLASSIFICATION OF AMBER FROM THO LOS A AT KAKOVATOS (Athens National Museum, inventory number 5688) Sample Spectrum Computer Sample Spectrum Computer Desig- Number Classifi- Desig- Number Classifi- nation cation nation cation Beads Unidentified Objects 1 1543 Baltic A 1534 Baltic 2 1544 Baltic B 1535 Baltic 3 1545 Baltic C 1536 Baltic 4 1546 Baltic D 1537 non-Baltic 5 1547 Baltic E 1538 Baltic 6 1548 Baltic F 1539 non-Baltic 7 1549 Baltic G 1540 Baltic 8 1550 Baltic H 1541 Baltic 9 1551 Baltic I 1542 non-Baltic 10 1552 Baltic Spacers 11 1553 non-Baltic? S 1529 Baltic Pendants T 1533 Baltic Q 1527 Baltic U 1530 Baltic R 1528 Baltic V 1531 Baltic 1 1568 Baltic W 1532 Baltic 2 1569 Baltic 1 1558 Baltic 3 1570 schraufite ? 2 1559 Baltic 4 1571 schraufite? 3 1560 Baltic 5 1572 Baltic 4 1561 Baltic 6 1573 Baltic 5 1562 Baltic 7 1574 Baltic 6 1563 Baltic 8 1575 schraufite? 7 1564 Baltic 9 1576 schraufite ? 8 1565 Baltic 10 1577 Baltic 9 1566 Baltic 10 1567 non-Baltic ? The shape of these apparently unique pendants is illustrated by Muller on Plate xv, 22 and here in Figure 2.1. The mounted exhibit in the National Museum includes the upper half of a very large 1 ; 2 3 -""iI/' ',,{ --', .. ;:1 ' ..... 1 ... , ........... / .. " ... / r, .. 1 t'\ ...... ",,,., f',,'1 ~ .. "J ~,,'I ~ .. ~,,,, , "., " 1 ' ,. , • ,. , , :. ,. ,I " II ., , : .' " " ., ". .' : • ,f ,I lid' ': t: I I .. " 11 , '. " , • f I ':' 2: .: I. t I I • • I t I ' • I I I I I I • I ' • I I ' , '. I , " • t..l~ :, ...... ~ :/ ..... ~ ',~J L/"'~ W"" • :/'\~ , ...... , /'.. .,...... /'... /"" ,,\"i ." .. • .. ~ "- ~ ,,- - j -, ~~-~-- •••• 6 4 Figure 2. AMBER PENDANT, COMPLEX BEADS, AND SPACERS FROM KAKOVATOS, THOLOS A ~ o ~ ~ ~ ~ ~ n tr:1 ~ ~ ~ ~ F ~ o ~ ..., o en CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 11 specimen having a maximum preserved width of 61.5mm and a smaller whole pendant which is 36mm wide and equally high. Neither were sampled since many identifiable fragments of these pendants were found in the storeroom. From these we have collected twelve samples whose partial spectra are shown in Figure 3. The com­ puter classification, summarized in Table 1, assigned eight of these spectra to Baltic amber. The other four spectra, nos. 1570, 1571, 1575, and 1576 differ markedly from those of Baltic amber in that the maximum absorption occurs at 8.2 [..I. with a gradual or stepwise decrease to 8.7 [..I.. W have never before found such a pattern in any sample of Baltic amber, no matter how badly deteriorated it was. 1527 1569 1572 1575 1528 1570 1573 1576 U68 .8 9 8 9 8 9 8 9· .1:' Figure 3. PARTIAL INFRARED-SPECTRA OF AMBER PENDANTS. KAKOVATOS, THOLOS A 12 MINOAN AND MYCENAEAN AMBER, III. KAKOV ATOS Nor can this unexpected result be laid to contamination, since none of the fragments tested seem to have received any conservation treat­ ment beyond cleaning. The pattern of these four samples is, in fact, the same which we have found in mineralogical specimens of schraufite, a fossil resin native to the Balkans and perhaps to the Lebanon. The computer, which has "seen" spectra of authentic schraufite, promptly classified the spectra of these four pendant frag­ ments as schraufite. The evidence is suggestive but far from conclu­ sive. It invites the notion that some of the pendants may have been made of fossil resins nearer to hand than the Baltic deposits, which in turn suggests local manufacture. But since the majority of the pen­ dants, as of all the other amber objects in Tholos A, are unquestion­ ably of Baltic amber, such an assignment, even if it is correct, does not alter the fact that Mycenaean Kakovatos participated in the amber trade in the Middle Bronze Age. Tholos A: Various Beads We return to Miiller's description:9 "There is further a tubular bead (Umm long, 6mm diam­ eter) and several curious pieces which look like combinations of simple beads. In three specimens two beads with beveled edges appear to be joined (Plate xv, 20; length 27 to ca. 45mm) [Figure 2.2]; as two of them have three drillholes, they may have served as links between a single and a double-stranded chain. The third with only two drillholes, as well as a piece which combines three flat beads (Plate xv, 19; length 26mm) [Figure 2.3], may have served to hold several chains together. A similar purpose, possibly to hold the strands of a tassel, can be suspected for a flat piece (Plate xv, 21; width 16mm, thickness 9mm, remaining length 25mm) [Figure 2.4]. It is semicircular at the preserved narrow side, and ten drill­ holes follow its rounded edge at irregular intervals." One of the double beads (Figure 2.2), the triple bead (Figure 2.3), and the fragment with ten holes (Figure 2.4) are now part of the mounted exhibit in the National Museum. Because of their unique­ ness, we have not taken samples from them. Instead we insert here the analyses of nine fragments of unidentified amber objects from the , Miiller, op.cit. (supra n.5) 280. CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 13 store rOOIll. Their partial spectra are shown in Figure 4. Computer classification, as summarized in Table 1, showed six of them to correspond to Baltic amber; the other three (spectra 1537, 1539 and 1542) did not meet the computer's requirements for Baltic amber, but 1.5'34 1537 1540 1535 1538 1541 1536 1539 1542 8 9 8 9 8 9 Figure 4. PARTIAL INFRARED-SPECTRA OF UNIDENTIFIED AMBER OBJECTS. KAKOVATOS, THO LOS A 14 MINOAN AND MYCENAEAN AMBER, m. KAKOVATOS inspection shows that, unlike the fragments of some of the pendants discussed above, they show no unusual features and are most likely only Baltic amber samples which have weathered beyond the point where computer classification of infrared spectra is possible. Tholos A: Spacers (Schieber) The most interesting amber objects from Tholos A are described thus by Muller :10 "Finally two series of ornaments deserve to be emphasized of which at least one example of each could be reconstructed reliably. The first is a trapezoid plate (7mm thick and 40mm high) with a [transverse] hole and nine [longitudinal] radiately arranged drillings of which the central one passes through the [transverse] hole. The other plate is rectangular, 75 by 38mm with a thickness of 9mm; both its principal surfaces slope slightly towards the narrow ends. It, too, has a [transverse] central hole and in addition 8 parallel [longi­ tudinal] drillings. So far it is reminiscent of the agate platelets [also found in Tholos A]. From the end of each drilling, however, there is a further drilling [actually two except for the first and last] approaching the neighboring drilling at an angle of 45° until it meets the [angular] drilling approaching from it. How were these curious pieces used? One can im­ mediately imagine strings passing through the many drillings, so that here, too, we are faced with members of a necklace. The central [transverse] holes could then have held precious stones to decorate the surface. Yet this explana­ tion does not account for the numerous oblique drillings along the edges of the rectangular pieces. These drillings may have had no purpose other than to reflect light and thus to enliven the [appearance of the] surface. That would of course be equally the case if strings had been passed through them, giving these ornaments a particularly splendid effect." In the later German literature these extraordinary objects (Figures 2.5 and 2.6) are called 'Schieber', a word which conveys little of their function in that language and less in translation. In English they are sometimes called 'spacer beads', but that term should be reserved for the small cylinder discs which are sometimes found separating 10 Miiller. op.cit. (supra n.5) 280-81. CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 15 larger and more or less spherical beads from one another in a neck­ lace. The 'Schieber' are beads only in the broad sense that they have drill holes for stringing; since their purpose was to separate the strands of a multiple necklace from one another, rather than the beads of a single strand, we call them simply 'spacers'. The spacers from Kakovatos are clearly related to similar ones from western and northern Europe. Merhartll first compared them with ten spacers from six tumuli of the South German Tumulus Culture (Hugelgraberkultur) and found them sufficiently alike to conclude, "that the spacers of Kakovatos are imports from the North will remain a certainty even if the amber used to make them should not be Baltic as is commonly assumed."12 Merhart did not extend his view to the amber spacers reported in England by Stukeley,13 Hoare,14 ThurnamI5 and others during the past two centuries. Much later, Milojcic considered the newly dis­ covered spacers of Grave 0 in Grace Circle B at Mycenae, which dates to about 1525 B.C.16 He acknowledged the implications of these finds to the dating of the South German Tumulus Culture, and he compared them as well to the spacers of the English Wessex culture, but concluded that only those from Lake, Wiltshire, correspond closely, while the rest are "helpless imitations." A different view of the relationships between the spacers from Greece, Germany and England has been taken by Hachmann.17 In an exhaustive investigation, he compared spacers from seven sites in England and forty sites in southern Germany (including Hesse, the middle Rhine, Alsace, Wiirttemberg and Bavaria) with the spacers from Kakovatos and Mycenae, and concluded that the Greek spacers derive from England rather than from South Germany and that the earliest Greek spacers are substantially older than the oldest datable spacers in southern Germany. The first conclusion is subject to some restrictions: Hachmann himself points outlS that the spacers from Kakovatos and Mycenae 11 G. von Merhart, Germania 24 (1940) 99-102. 12 von Merhart, op.cit. (supra n.ll) 101. 13 W. Stukeley, Stonehenge (London 1740) 45 and Plate XXXII. 14 R. C. Hoare, The Ancient History of Wiltshire (London 1812-21). 16 ]. Thurnam, Archaeologia 43 (1871) 500ff. 16 V. Milojcic, Germania 33 (1955) 316-19. 17 R. Hachmann, Bayerische Vorgeschichtsbliitter 22 (1957) 1-36. 18 Hachmann, op.cit. (supra n.17) 13. 16 MINOAN AND MYCENAEAN AMBER, m. KAKOVATOS correspond closely not only to some spacers from England (Upton Lovell,19 Oakley Down,20 and Huntiscarth (Knowes of Trotty),21 but also quite as closely to some South German spacers (Hundersingen­ Weidenhang,22 Mehrstetten,23 Melchingen-Hakenrain,24 Upflamor­ Lautrieb,25 and Wilsingen-Stockacker26). The spacer from Koblach, Vorarlberg, Austria, should be included in this group.27 Hachmann's relative dating of the German spacers is of necessity limited and vague. As Miss Sandars has pointed out,28 the fact that the first ap­ pearance of such spacers in Greece must now be dated to 1600 B.C. or even earlier does not make it necessary to try to date the German specimens to an equally early period; the evidence as reviewed recently by Professor Piggott29 shows "that Aegean, and largely Mycenaean, interest in the west and northwest was not an episode of short duration, but something continuous throughout the whole of the Late Helladic period." The decision whether the Mycenaean amber spacers with complex borings correspond more closely to those of the English Wessex cul­ ture or to those of the South German tumulus culture cannot be made with any finality. Moreover, one may doubt whether it is profitable to champion either relationship. The more important fact is that these spacers are obviously all related, and that we can now add amber spacers from Andrup near Esbjerg on the west coast of Jutland.30 Seen as a whole the distribution of these spacers supports a view of Mycenaean connections with the north of Europe which is now gaining ground :31 sea-traffic from Greece to the western Mediter- 1. W. Cunnington, Archaeologia 15 (1806) 127 or 129; Thurnam, op.cit. (supra n.15) 501; Hachmann,op.cit. (supra n.17) fig. 12:45,46. 10 Hoare, op.at. (supra n.14) 238. 11 S. Piggott, ProcPS 4 (1938) fig. 13A; Hachmann, op.cit. (supra n.17) fig. 2. IZ G. Kraft, Die Kultur der Bronzezeit in Suddeutschland (Augsburg 1926) 123 no. 026; Hachmann, op.at. (supra n.17) 34 and fig. 11 :5-13. 13 Kraft, op.cit. (supra n.22) 125 no. Ell; Hachmann, op.cit. (supra n.17) 34 and fig. 9: 16. 24 Kraft, op.cit. (supra n.21) 117 no. B21; Hachmann, op.cit. (supra n.17) 35 and fig. 9:30. 16 Kraft, op.at. (supra n.22) 19 no. C8,5; Hachmann, op.at. (supra n.17) and fig. 11: 15,21. 18 Kraft, op.cit. (supra n.22) 120 no. 24,2; Hachmann, op.at. (supra n.17) 35 and fig. 10: 25,26. 17 N. K. Sandars, Antiquity 33 (1959) 292-95, does so include it. 18 Sandars, op.at. (supra n.27) 294. 18 S. Piggott, Sbornik Narodniho Muzea v Praze, SER. A, 22 (1966) 117-25. 80 E. Lomberg, Antiquity 41 (1967) 221-23. 31 E. G. J. Bouzek, Pamdtky Archeologiclee 57 (1966) 242-76; Piggott, op.cit. (supra n.29); K. Randsborg, Acta Archaeologica 38 (1967) 1-27. CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 17 ranean, whence an overland route along the Rhone-Rhine valleys leads to England and Scandinavia. The South German spacers then could have emanated from Alsace eastwards ;32 although a separate (and possibly later) connection between Jutland and southern Ger­ many is an alternative possibility that may help to solve the problems in the chronology of the German spacers. But the evidence for a land route across eastern Europe is by no means negligible. Finds of Mycenaean origin or affinity, as sum­ marized most recently by Sulimirski,33 leave little doubt that there was a lively traffic between the Aegean and the Danube Basin be­ tween the eighteenth and thirteenth centuries B.C. Whether the con­ centration of such finds in Denmark during this period is necessarily a direct extension of this thrust remains a question to which Suli­ mirski's map gives no clear answer. But it would be most unreason­ able to deny that some exchange will likely have taken place across the narrow and well-populated gap which remains between central Europe and Jutland. The choice between a western sea route and an eastern land route is therefore one it might be wiser to reject than to make. There is no reason why the two should be mutually exclusive. There is an analogy to such a solution in Hencken's compelling analysis of the distribution of the Herzsprung shields oflater times.34 He showed that the V-type shields spread from the Aegean by a western sea route via Spain to Ireland, while the U-type shields traveled by an overland route from the head of the Adriatic to Denmark and thence to Ireland. This overland route is, of course, the main Iron Age amber route; Navarr035 has shown that it was of little if any consequence until about 1000 B.C., and thus it cannot be drawn upon to map the Myce­ naean amber trade. To repeat, we are not suggesting that Hencken's Herzsprung paper describes the trade routes of a period which lies 600 to 800 years farther in the past, but we do think that it provides the model for the 32 F. A. Schaeffer's opinion (Les Tertres funeraires prehistoriques dans la Foret de Haguenau I [Haguenau 1926] 262) that the Alsatian amber is likely of Italian provenance is based on the most tenuous chemical evidence and should be ignored until a spectral analysis of these finds has been made. 33 T. Sulimirski, in Teki Historyczne (London 1966-68) 3-27. 34 H. Hencken, A]A 54 (1950) 295-309. 35 J. M. de Navarro, Geogr] 66 (1925) 481-507. 2-G.R.B.S. 18 MINOAN AND MYCENAEAN AMBER, ill. KAKOVATOS kind of pincer movement which seems to fit the evidence for the amber trade in Late Mycenaean times. The western prong would have reached across the length of the Mediterranean and then across France to Wessex and ultimately to Jutland. The eastern prong would have led up the Danube and reached Denmark across Germany. Whether the two routes were entirely synchronous, overlapped, or succeeded each other is a question which would merit further exam­ ination. It is tempting to hope that the problems which have arisen from the amber spacers found in France, South Germany, Austria and Bohemia could be much reduced by considering that some of them may be offshoots of the western route while others emanate from the eastern one. We return to the simpler matter of analyzing the amber of the Kakovatos spacers.36 Two large fragments are mounted on the board for exhibition in the National Museum. No samples have been taken from these. Five more large pieces were found in storage. They are listed in Table 1 as S, T, U, V and W. S is a roughly rectangular frag­ ment (height 32mm; width 20mm; thickness 9mm) which retains one complete boring and one along each edge of fracture. T is a trapezoidal end-spacer (height 23mm; maximum width 21mm; thick­ ness only 5mm) which appears essentially complete. Since it is smaller and has only four radiately arranged borings, it cannot be the end­ piece described and pictured by Muller or even have likely been part of the same assembly. U is also apparently complete, rectangular (height 27.5mm; width 20mm; thickness 8.5mm), and with three parallel borings. V is a large fragment of a rectangular piece (height 32.5mm; width 30mm; thickness 9mm) with three parallel borings. W is another fragment of a rectangular plate (height 23mm; width 23mm; thickness 9mm) with two intact parallel borings and one along the edge of fracture. Samples from these five spacers and from ten unmarked small fragments of spacers were analyzed. As shown in Figure 5 and Table 1, fourteen of these samples give spectra which the computer classified as those of Baltic amber; the exception, spectrum 1567 of spacer sample 10, differs only very slightly from the others and fails the computer test by a hair. 8e spectral analyses of the spacers from Grave Omicron, Circle B. of Mycenae will be reported in the next paper of this series. CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 19 Tholos B MulIerS' lists only" one small bead of amber," but gives neither a description nor an illustration. We have not been able to find it in the National Museum at Athens. For a possible but unverifiable identification of this bead, see infra. Figure 5. PARTIAL INFRARED-SPECTRA OF AMBER SPACERS. KAKOVATOS, THO LOS A Tholos C Muller38 describes "seven small and one larger drilled amber beads in the form of very flat plates. Diameters 8-12mm and 20mm, respectively. " Mounted on a small board in the National Museum, together with seven amethyst beads, a plane-convex disc of rock crystal, a gold but­ ton and fragments of gold leaf and bronze, there are four amber beads. Three of them are more or less uniform in shape and size: 37 Miiller, op.cit. (supra n.5) 295. 38 Miiller, op.cit. (supra n.5) 301. 20 MINOAN AND MYCENAEAN AMBER, ill. KAKOVATOS cylinder discs with diameters of 8-10mm. The fourth is a distinctly oval drilled bead with a short axis of 17mm and a long axis of 20mm. Found unmounted in a box labeled HTholos C" were another four amber beads and some unidentifiable fragments. Three of the beads are again cylinder discs with diameters of 9-11mm, quite like the three mounted beads. The fourth unmounted bead has a diameter of 20mm and is somewhat irregular in shape, rather suggestive of an octagon with alternately long and short sides. We have thus six small and two large beads, rather than seven small and one large bead as listed by Muller. It is possible that one of the small beads has been broken into the unidentifiable fragments found. The unlisted larger bead may then actually be the one which Muller described from Tholos B. We have analyzed samples from each of the four unmounted beads. These are listed as samples A-D in Table 2. Eight of the un­ identified fragments are also listed in this table as Samples AI_HI. Partial spectra of all twelve samples are shown in Figure 6. Computer classification identifies eight of the twelve spectra as those of Baltic amber. Of the exceptions, spectrum 1516 shows extensive deteriora­ tion, spectrum 1519 shows reverse absorption like spectra 1570 and 1571 in Figure 3, and spectrum 1520 fails the test because of a minutely positive slope between 8.2 and 8.4 tJ., but is otherwise visually recog­ nizable as a typical Baltic amber spectrum. TABLE 2. SPECTROSCOPIC CLASSIFICATION OF AMBER FROM THOLOS C AT KAKOVATOS Sample Spectrum Computer Designation Number Classification A 1515 Baltic B 1516 non-Baltic C 1517 Baltic 0 1518 unclassifiable Ai 1519 schraufite? Bl 1520 non-Baltic ? Cl 1521 Baltic 0 1 1522 Baltic El 1523 Baltic Fl 1524 Baltic Gl 1525 Baltic Hi 1526 Baltic CURT W. BECK, CONSTANCE A. FELLOWS, AUDREY B. ADAMS 21 Figure 6 contains one partial spectrum which is wildly erratic: spectrum 1518 shows intense, broad absorption between 9 and 1O!.l. which completely destroys any resemblance to any fossil resin spec­ trum. This is probably caused by a major inorganic impurity, i.e., dirt. The computer, precisely following its program written for resin spectra, was unable to cope with this glaring difference and pro­ nounced the spectrum as one of Baltic amber. This is an instructive example of the obvious but easily forgotten limitations of the com­ puter approach: the gain in objectivity is of necessity a loss of judge­ ment, and an amber spectrum should not be entrusted to the com­ puter without a previous check for gross impurities. 1515 1518 1521 1516 1519 1522 1525 1517 1520 1513 Figure 6. PARTIAL INFRARED-SPECTRA OF AMBER BEADS. KAKOVATOS, THOLOS C 22 MINOAN AND MYCENAEAN AMBER, ill. KAKOVATOS Conclusion Of the 59 samples of amber from the tholoi of Kakovatos, 46 have been identified as Baltic amber, five (spectra 1570, 1571, 1575, 1576 and 1519) may be schraufite of Balkan or eastern Mediterranean origin, and eight are unidentifiable because of extensive deterioration, though three of these are so similar to spectra of Baltic amber that -their failure to pass the computer test is not significant. This establishes that at least the raw material of most if not all of the amber from Kakovatos must have been imported. The amber can only have come from northern Europe, since the number and size rules out southern Russia, where occasional pieces of Baltic amber occur naturally.39 Our findings thus agree with those of Jonas,40 who had analyzed one of the Kakovatos beads by an indirect and questionable41 adaptation of the succinic acid test. VASSAR COLLEGB January, 1970 ACKNOWLEDGBMENTS This work has been supported by U.S. National Science Foundation Grants GS-739 (Anthropology) and GP-4729 (Geochemistry), as well as by travel grant 729 (Johnson Fund) of the American Philosophical Society. The collection of samples in Athens was made possible by the assistance and co-operation of Professor H. S. Robinson of the American School of Classical Studies at Athens, Dr j. Kontis of the Greek Archaeological Service, and Dr V. Kallipolitis, Director of the National Museum. We thank Professor Christine M. Havelock of Vassar College for advice and criticism and Professor Stuart Piggott of the University of Edinburgh for several important references to the literature . . 39 C. W. Beck, GRBS 7 (1966) 209-11. 40 R. Jonas, Schrifun der Physikalisclt-Okoncmischen Gesellschaft {u Konigsberg 49 (1908) 351-68. u Beck, op.cit. (supra 0.39) 197-200.