Analysis and Provenience of Minoan and Mycenaean Amber, II. Tiryns Beck, Curt W;Southard,
Gretchen C;Adams, Audrey B Greek, Roman and Byzantine Studies; Spring 1968; 9, 1; ProQuest pg. 5

Analysis and Provenience of Minoan 
and Mycenaean Amber, II. Tiryns 

Curt W. Beck, Gretchen C. Southard, Audrey B. Adams 

THE GRAVE ROBBERS' HOARD known as the Tiryns Treasure in­
cludes two well-preserved and two (?) fragmentary gold-and­
amber objects which have been fully described but as yet not 

satisfactorily explained. The hoard was found in December, 1915, in 
the area of the Mycenaean lower town southeast of the citadel of 
Tiryns. Karol first thought that it had been buried in Geometric 
times but later2 changed his view in favor of burial "before the fall 
of the Mycenaean culture (vor dem Untergang der mykenischen 
Kultur)." It is now generally accepted that the objects in the hoard 
range widely within the Late Helladic period.s 

We are here concerned only with the gold-and-amber objects. 
Kar04 described them as "a pair of curious, wheel-shaped ornaments 
of goldwire, the spokes of which are set with amber beads." He added 
that the hoard yielded sufficient additional gold wire and amber beads 
to account for a second pair, and suggested that they may have served 
as "Schlafenschmuck" like the wheels covering the ears of the lady 
of Elche.5 In his first note, Karo6 supplied a photograph giving an 
axial view of one "wheel" containing seven amber beads (now Athens 
National Museum no. 6217b) and a side view of the other (now 
Athens National Museum no. 6217a) in which the amber cannot be 
seen clearly. 

In a somewhat later note, Philadelpheus7 described these objects as 
"three crowns (SuxS1}fLct'Tct) of woven gold wire ... containing crossed 
bronze rods in which large, rounded amber pieces are mounted"; he 
gave the number as three: two virtually intact and one of which only 

1 G. Karo, AA 1916, 143-147. 
2 G. Karo, AthMitt 55 (1930) 119-140. 
3 H. W. Catling, ProcPS 22 (1956) 111. 
« Karo, supra n.l. 
5 P. Paris, MonPiot 4 (1897) plate 13/14. 
6 Karo, supra n.l. 
7 A. Philadelpheus, De/tion 2 (1917) Suppl., 17. 

5 



6 ANALYSIS AND PROVENIENCE OF AMBER, II 

the amber pieces were found, although he listed two balls of gold 
wire (now Athens National Museum no. 6218). His illustrations show 
no. 6217a (with five amber beads) and no. 6217b (with seven amber 
beads); the third "crown" is represented by seven amber beads 
mounted on two modem metal rods, but without any gold wire. This 
is now Athens National Museum no. 6219. 

In his full publication of the Tiryns Treasure, Karos included 
photographs of the intact "crowns" in side views (Beilage xxx A, re­
produced as our PLATE 1) and in axial views (Beilage XXXI). He gave 

No. 6217A No. 6217B 

Figure 1. THE Two INTACT GOLD-AND-AMBER ORNAMENTS OF THE TIRYNS HOARD 

measurements of the amber beads but no designations for them. We 
have numbered the beads with Roman numerals (in the order in 
which Karo listed them) as shown in Figure 1 and in Table 1.9 Karo 
also illustrated the loose amber pieces mounted on their modern rods 
(Beilage XXXII); Figure 2 has been traced from his photograph. It 
shows, as did Philadelpheus' illustration,Io seven pieces of amber: 
four mounted on one rod and three on the other. Again we have 
numbered these with Roman numerals in Figure 2 and in Table 1 in 
the order in which Karo listed them. 

8 Karo, supra n.2. 
I Karo's and our measurements do not agree exactly in all cases (cf. Table 1), but they 

agree closely enough for identification. We will not claim that our measurements are more 
accurate than Karo's, although they were taken with slide calipers; many of the amber 
beads, closely mounted on their bronze rods, are not easily accessible. There may also 
have been a reduction of size by breakage, or an increase by the application of protective 
coatings. 

10 Philadelpheus, supra n.7. 



c. W. BECK, G. C. SOUTHARD, A. B. ADAMS 7 

TABLE 1. DIMENSIONS AND SPECTROSCOPIC CLASSIFICATION OF AMBER BEADS 
IN TmYNs HOARD 

Athens Maximum Dimensions (mm) 
National Desig- Karo Authors Spectrum Extraneous Computer 
Museum nation L D L D Number Absorption Classifica-
Number tion 

6217 a-I 26 23 28 20.5 782 7.81-' Baltic 
a-II 25 18 28 20 785 7.8 P. (weak) Baltic 
a-m 24 15 25.5 15 
a-IV 19 14 18.5 15 809 none Baltic 
a-v 39 22 42 15 786 7.8 P. Baltic 

and 
790 

b-I 26 33 25 33 
b-II 19 29 19.5 22.5 783 7.8 p. Baltic 
b-m 24 24 25 23.5 
b-IV 24 20 24.5 19.5 784 7.8 P. Baltic 

791 7.25 and 7.8 p. Baltic 
b-v 12 26 14 28.5 
b-VI 27 17 29 16 808 none Baltic 
b-VII 11 9 11.5 10 781 7.25 and 7.8 p. Baltic 

6219 I 31 34 31 34 787 7.81-' Baltic 
803 none Baltic 

II 27 36 29 36.5 799 7.25 and 7.8 p. Baltic 
804 none Baltic 

m 20 37 19.5 36.5 806 7.25 and 7.81-' Baltic 
IV 20 31 18 32 807 7.8 P. (weak) Baltic 
v 17 15 22 15 
VI 21 11 22.5 11 788 7.8 P. (weak) Baltic 
VII 13 20 no longer in collection 

not listed 13.5 11.5 
unnumbered fragment 805 none Baltic 
unnumbered fragment 801 7.25 and 7.81-' Baltic 
unnumbered fragment 800 none Baltic 

However, we find only six amber beads mounted on the two 

modem rods in the National Museum, three on each. The seventh 
bead listed by Karo with a length of 13 mm and a diameter of 20 mm 
no longer exists. There is an unmounted bead, length 13.5 mm, dia-
meter 11.5 mm, which cannot be Karo's No. 6219-vn and which is so 
badly weathered that it has been impossible to take a usable sample. 



8 ANALYSIS AND PROVENIENCE OF AMBER, II 

In addition, there were eight fragments which may well be what 
remains of Karo's No. 6219-vn and of which we have sampled three 
as shown in Table 1. 

No. 6219 

Figure 2. AMBER BEADS FROM THE TIRYNS HOARD MOUNTED ON MODERN RODS 

Sample Contanlination 

We have taken samples of thirteen of the eighteen whole mounted 
beads and of three of the fragments, for a total of sixteen samples. 
The sampling, never an easy matter with archaeological amber arti­
facts, was very much complicated by thick coatings of foreign mater­
ials. Doubtless because of their uniqueness, the amber of the Tiryns 
wheels has received rather more conservation treatment than has 
been good for it for analytical purposes. One heavy coating, which 
completely fills many of the cracks and drill holes of some pieces, has 
a waxy consistency. A sample of it gave the unmistakable infrared 
spectrum of beeswax.ll Inquiries at the National Museum yielded no 
information on this or other treatments; they were evidently made 
before records of conservation work were kept. A second treatment 
with a synthetic resin seems to have been made; again there is no 
record, and the resin now used at the National Museum, a com­
mercial fingernail polish, is not the same as that contaminating the 
Tiryns amber, as the infrared spectra clearly show. 

As had to be expected, these treatments produced extraneous bands 
in the infrared spectra of the amber. One of these, a band at about 

11 W. Kuhn, Studies in Conservation 5 (1960) 74. 



" 

C. VV. BECK, G. C. SOUTHARD,A. B. ADAMS 9 

7.8fL (1280 em -1), due to beeswax, is perilously close to the 8 to 9fL 

(1250 to 1110 em -1) region which we have found to be most useful for 
the identification of Baltic amber.12 Where enough sample was avail­
able for duplicate tests (samples 6217b-IV, 6219-1, and 6219-n), the 
inner surface of the sample chip which ought to be free (or freer) of 
contamination gave the better spectrum; indeed, in the two latter 
cases, it gave a perfect one. The second impurity is less prevalent. 
Five of the nineteen spectra run show a curious sharpening of the 

A B C D 

Figure 3. EFFECT OF IMPURfTIES ON THE 7.25 fL (1380 em-I) METHYL BAND OF 

AMBER 

A. Methyl band of amber. B. Methyl band of hypothetical synthetic resin. c. Super­
position of A and B. D. Methyl absorption band of contaminated Tiryns amber 

symmetrical methyl absorption at 7.25 fL (1380 em -1) as shown in 
Figure 3d. We have found a similar absorption pattern in the spectra 
of the surface, but not of the interior, of a Near Eastern amber object 
and we believe that the following considerations offer an explanation 
for it. Although the complete chemical composition of amber (and 
of all other fossil resins) is not yet known, all these resins give infrared 
spectra which leave no doubt that they contain several methyl groups 
in different molecular environments. Because of these slight differ­
ences in environment, the methyl groups absorb infrared radiation 

11 C. W. Beck, GRBS 7 (1966) 191-211 and earlier work cited there. 



10 ANALYSIS AND PROVENIENCE OF AMBER, II 

over a range of wavelengths near 7.25 p. (1380 cm -1) leading to a 
typically broad and rounded absorption peak as shown for Baltic 
amber in Figure 3a. Pure compounds, on the other hand, if they con­
tain only one methyl group per molecule, or if they contain more 
than one methyl group but in essentially identical environments, 
show very sharp absorption bands in this region, as in Figure 3b. This 
is precisely what would be expected from the theory of infrared 
spectroscopy, since only identically bonded methyl groups will require 
identical amounts of energy to support the vibrations which cause 
absorption of infrared light. Now if a sample contained a fossil resin 
with a broad methyl absorption band and, at the same time, a pure 
compound with a sharp absorption peak at the same wavelength, but 
of greater relative intensity, the absorption pattern would be a super­
position of Figures 3a and 3b as shown in Figure 3c. This is precisely the 
shape of the 7.25 p. (1380 cm -1) absorption in five of the spectra of the 
Tiryns amber samples, and we conclude that these samples have been 
treated with a compound containing a large number of identical 
methyl groups. This criterion is met by synthetic resins. While it is 
not possible to identify the contaminant from the sparse spectral evi­
dence, a synthetic lacquer of the vinyl acetate type is a very reasonable 
possibility.13 We have tried to remove these impurities with organic 
solvents in order to gain a H clean" amber sample which would give 
an infrared spectrum free of all extraneous bands. Amber itself is not 
completely soluble in any known material.14 To our disappointment, 
the contaminants proved to be exceedingly tenacious. Extraction with 
carbon tetrachloride, petroleum ether and ethyl acetate, in the cold 
or after boiling for as long as 30 minutes, failed to remove them, 
although it did reduce the intensity of the absorption bands which we 
ascribe to them. This in itself is good evidence that these bands are not 
caused by the amber, and we have therefore interpreted the spectra 
by ignoring them. As shown in Table 1, eight spectra have the band 
at 7.8p. (1280cm-1) which we assign to beeswax; five have both the 
sharp band at 7.25 p. (1380 cm -1), probably due to a synthetic varnish 
and the 7.8p. (1380 em-I) band of beeswax; only six spectra are en­
tirely free from extraneous bands. 

Because of the problems arising from such impurities we would 
13 Chicago Society for Paint Technology, "Infrared Spectroscopy" (Chicago 1961) Spec­

trum no. 23. 
14 L. Schmid, "Bernstein," in C. Doelter and H. Leitmeier, Handbuch der Mineralchemie 

IV.3 (Dresden and Leipzig 1931) 892. 



c. W. BECK, G. C. SOUTHARD, A. B. ADAMS 11 

like to exhort archaeologists to avoid treating amber artifacts in any 
manner unless a piece cannot be taken from the site without consoli­
dation. In the very few instances where that is the case, we would 
recommend purified paraffin wax in preference to beeswax. Paraffin 
is quite as difficult to remove15 as beeswax, but it has fewer absorp­
tion bands in the infrared because it contains only carbon-carbon and 
carbon-hydrogen bonds, and these do not absorb infrared radiation in 
the range of 8 to 9 J.L (1250 to 1110 em -1) which we have found most 
useful for the identification of fossil resins. Synthetic varnishes should 
be avoided because they are notoriously difficult to remove and be­
cause they give complex infrared spectra. 

Provenience of the Tiryns Amber 

The infrared spectra of the 16 samples listed in Table 1 are shown 
in Figure 4 over the range of 8 to 9 J.L (1250 to 1110 em -1). Where two 
spectra were made of a given sample, the better one is shown. The 
slope of the absorption in this range is the most useful criterion for 
the recognition of Baltic amber. In a well-preserved specimen it has a 
value of zero between 8.0 and 8.5 J.L (1250 and 1175 em -1), i.e. there is a 
horizontal line or <C shoulder" ;16 absorption then increases and reaches 
a maximum value at 8.7 J.L (1150 em -1).17 In archaeological samples 
this pattern is almost invariably changed in varying degrees by 
atmospheric oxidation resulting in continuously increasing absorption 
over the entire range from 8.0 to 8.7 J.L (1250 to 1150 em -1).18 This 
oxidative effect is most pronounced in spectra 800,805 (both of sample 
6219; unnumbered fragments), 807 (sample 6219-IV) and 808 (sample 
6217b-VI). In most samples, however, the horizontal shoulder is well 
preserved, most markedly so in spectra 804 (sample 6219-rr), 782 
(sample 6217a-I) and 788 (sample 6219-VI). In those spectra which have 
a strong extraneous beeswax peak near 7.8 J.L (1280 em -1), the slope of 
the absorption pattern near 8 J.L (1250 em -1) is affected with the result 
that the slope assumes a positive value between 8.0 and 8.2 J.L (1250 
and 1220 em -1). Spectrum 784 (sample 6217b-IV) is an example. 

15 M. Wheeler, Archaeology from the Earth (Penguin 1961) 197. 
16 Beck, supra n.12; C. w. Beck et a!., "The Infrared Spectra of Amber and the Identifica­

tion of Baltic Amber," Archaeometry 8 (1965) 96-109. 
17 Since infrared spectra are normally plotted with zero absorbence at the top of the 

scale, a maximum absorption will deflect the recording pen downward so that the curve 
appears to have a minimum if read from the bottom. 

18 Beck, supra n.12, figs. 1 and 2. 



12 ANALYSIS AND PROVENIENCE OF AMBER, IT 

785 786 788 800 

801 803 804 805 

H06 807 808 809 

Figure 4. INFRARED SPECTRA OF THE TIRYNS AMBER FROM 7.5 TO 9 r- (1330 TO 

1110 em-I) 

Computer Classification of Amber Spectra 
The distinguishing features of the infrared spectra of fossil resins, 

and especially of archaeological amber artifacts, are subtle; classifica­
tion more often relies on variations of slope than on the presence or 
absence of strong, sharp absorption bands. We have been aware that 
the interpretation of these spectra calls for a practiced eye and we are 
sensitive to the charge that a practiced eye may be merely a euphem­
ism for a subjective view. In order to reduce the personal element 



c. W. BECK, G. C. SOUTHARD, A. B. ADAMS 13 

from the determination of the provenience of amber by infrared 
spectroscopy, and in order to compare new spectra quickly to well 
over a thousand reference spectra, we have programmed an IBM 360 
Model 30 Computer to analyze and classify the infrared spectra of 
fossil resins. After trials with a variety of numerical inputs read from 
the spectral curves, we have found that a rather simple mathematical 
analysis of the slopes between 8 and 9"" (1250 and 1110 em -1) serves 
well to characterize Baltic amber and to distinguish it from all the 
other fossil resins we have tested so far. This range also gives promise 
to permit the absolute identification of non-Baltic resins, e.g. simetite 
from Sicily, schraufite from the Bukowina, walchowite from Moravia, 
etc., but it will require more samples and more work to reach the 
point where the computer can identify these resins with reasonable 
certainty. At this time we have only a program which allows the com­
puter to identify all but two of 60 samples of Baltic amber and to 

reject all but one of the 58 samples of non-Baltic amber which we 
have tested by using geological samples of known origin. 

The program is simple enough.19 The computer is given intensity 
readings at 0.1"" intervals between 8.0 and 9.0 "". After standardizing 
these readings by dividing them by the intensity at 8.0"", it then 
computes simplified slopes and their second derivatives over selected 
0.1 and 0.2"" intervals.20 Using a combination of four of these functions 
as criteria, the computer classified all but three of 118 samples, or 
97.5%, accurately as Baltic or non-Baltic. 

Since computers, like women, are more widely admired than 
understood, we hasten to add that we do not see in this approach 
more than a rapid and convenient way of doing what could be done 
equally well, though much more laboriously, with pencil and paper. 
The computer does not select the intensity readings to be used or the 
functions to be calculated; we selected these as the most useful ones 
after much trial and error which would not have been practical or 
economical without using the computer. The program merely com­
pares four functions from one spectrum to four corresponding limits 
determined by the function values of a large body of spectra of Baltic 
and non-Baltic amber of known origin. We are satisfied that any 
sample which agrees in these respects with the substantial number of 

It Details of the programming procedure will be published elsewhere. 
110 The term slope is used loosely. The computer actually finds the differences between 

the standardized intensity values. A slope, sensu stricto, is the ratio between two differences 
on the abscissa and the ordinate. 



14 ANALYSIS AND PROVENIENCE OF AMBER, II 

samples of authentic Baltic amber does in fact come from the area of 
natural distribution21 of this species. 

By these criteria, the computer classification of all the 19 spectra of 
the Tiryns amber is that of Baltic amber. 

Origin of the Tiryns "Wheels" 

Our proper task is the provenience analysis of archaeological amber 
artifacts. The uniqueness of the Tiryns «wheels" may excuse some 
additional comments. 

Our analyses establish firmly that the amber of these remarkable 
ornaments is of the Baltic variety. It does not reveal the place where 
the "wheels," or even the amber pieces in them, were fashioned into 
their present form. The shape of the amber beads is one of the enig­
mata which these extraordinary objects present to the archaeologist. 
Kar022 described them as having «eine doppelt geschweifte Form, die 
am ehesten an ein kurzes Stuck Rohr mit Knoten erinnert." There 
are a number of Mycenaean amber beads of the same curious shape. 
One is Ashmolean Museum no. AE 1515 from the Dictaean (Psychro) 
cave, which Boardman23 describes as "barrel-shaped." A second, un­
registered bead in the Ashmolean, from the same source, has the 
same shape, though in a more elongated form. Another was listed by 
Evans24 among the trinkets passed with the Thisbe Treasure; the 
only amber piece from this source now in the Ashmolean Museum 
has lost all recognizable shape. Evans suggested an LM lA date for it. 
The British Museum has two further examples from Tomb 13 at 
Ialysos on Rhodes (Strong25 nos. Ib and lc). The pottery from this 
tomb is all LH me. Strong calls them" double concave."26 The island 
of Kephallenia has yielded a substantial collection, both at Lekkithra27 

and especially at Metaxata.28 A single example from Salamis, Athens 
National Museum no. 3589, was described by Wide29 as a "tubular 

II Baltic amber is not limited to the Baltic Sea but occurs widely across northern Europe 
from England to deep into Russia. 

n Karo, supra n.2. 
II J. Boardman, The Cretan Collection in Oxford (Oxford 1961) no. 352, p.73; fig. 32, p.72; 

plate XXllI. 

U A. Evans, "The Ring of Nestor," JHS 45 (1925) 2. 
IIIi D. E. Strong, Catalogue of the Carved Amber in the Department of Greek and Roman Antiqui-

ties (British Museum, London 1966). 
16 Strong, supra n.25, p.39. 
17 S. Marinatos, ArchEph 1932, p.42 n.2 and plate 15, A3 and A7. 
28 S. Marinatos, ArchEph 1933, pp.92-94 and plate 2. 
28 S. Wide, AthMitt 35 (1910) 31. 



C. ~. BECK, G. C. SOUTHARD, A. B. ADAMS 15 

bead of bone(?)," but Mr Carling in Oxford has examined it and finds 
it to be amber.30 Far afield there is a necklace of 18 large, more or less 
well-preserved beads of this shape from Tomb 31 of the necropolis of 
Piazza Monfalcone on Lipari, dated to the first phase of the Ausonian 
II period, or 1150 to 1050 B.C.31 They have been described as "of a shape 
which sometimes recalls that of astragaloi."3 2 Perhaps there are 
others. It may be significant that no beads of this shape seem to be 
known in any material other than amber. 

It is also notable that all the beads of this type which can be dated 
fall within a narrow range of time near 1200 B.C. Evans LM IA date is 
not well founded; the Thisbe trinkets are now considered to span the 
range of LH I to LH m. 

There is obviously no consensus about their shape. From the pub­
lished illustrations and from direct examination we have no doubt 
that an organic, not a geometric, representation was intended: the 
cross-section of the beads is almost invariably elliptical rather than 
circular, the ends are usually not parallel, and the central ridge is not 
placed by any considerations of symmetry. Though we cannot quite 
follow Brea and Cavalier33 in seeing these beads as astragaloi, they do 
resemble bones more than anything else that comes readily to mind. 
If they are bones they might, like astragaloi, be good luck symbols or 
they might have a more specific and religious significance as the rem­
nants of sacrificial animals. 

The technique of the gold wire work of the Tiryns "wheels" is 
unique in Mycenaean archaeology.34 A length of gold wire, of circular 
cross section and 1 mm diameter, has been doubled back upon itself 
and coiled ten or eleven times into a circle which thus has 20 to 
22 horizontal turns of wire. These are connected by half-round gold 
wires which form alternately vertical lines and rhomboid patterns. 
The half-round wire is wrapped once around each turn of the 
horizontal wire in passing; in the vertical connections the direction of 
these turns is the reverse of that in the rhomboid ones (Figure 5). 
Kar03s called the technique Hin imitation of weaving (der Weberei 

aD H. W. Catling, by personal communication. 
al L. BernabO Brea and M. Cavalier, Meligunis-Lipdra I (Palermo 1960) 119 and plate 

xLm-l. 
31 Brea and Cavalier, supra n.27, p.149. 
33 Brea and Cavalier, supra n.27, p.149. 
34 Karo, supra n.2. 
35 Karo, supra n.2. 



16 ANALYSIS AND PROVENIENCE OF AMBER, II 

Figure 5. WRAPPING OF RHOMBOID AND VERTICAL CROSS-WIRES IN THE GOLD­

AND-AMBER ORNAMENTS OF THE TmYNs HOARD 

nachgebildet)"; Marinatos36 refers to it as Hplaited (geflochten)." In 
weaving it is known as the wrapped or Soumak pattem;37 in basketry 
it is one of the coiled-work methods called bee skep.38 That the 
objects imitate basketry is also suggested by the combination of 
rounds (called rods or sticks in basket weaving) with half-rounds for 
cross-ties (called skeins in basket weaving and made by splitting young 
growths longitudinally). Surely the goldsmith had a reason for going 

so s. Marinatos, "Lausitzer Goldschmuck in Tiryns," Deutsche Beitritge ~ur A/tertums­
wissenschaft 12/13 (1960) 151-157. 

81 H. Hodges, Artifacts (New York and London 1964) 143, fig. 36-3. 
88 R. J. Forbes, Studies in Ancient Technology IV (Leiden 1956) 174 fig. 18; 186 fig. 26. Both 

the word (which means bee hive) and the technique are held to be of Scottish origin. 



BECK, SOUHL\RD, AlH~lS PL\ TE 1 

GOL[)-A.'\D-A~IBER OR:\,AME:'\TS FRO:'1 THE Tnu:'\s H(LUW 

AnIE:\s ~.HIOX\L i\kSEDl ;'\os. 6217.'. Clipper) A:\D 6217B (io"w) 



c. W. BECK, G. C. SOUTHARD, A. B. ADAMS 17 

to the considerable trouble of shaping the gold wire for the vertical 
connections into a semi-drcular cross section. 

Philadelpheus' appellation <Ccrowns"39 never sat well with the 
Tiryns ornaments. Karo's40 name "wheels" was obviously suggested 
by the cross rods which he thought of as spokes, but the fragility of the 
"wheels" and the absence of any hub or other means to connect them 
to an axle speak against the idea. Were they, then, perhaps baskets? 
They have of course no bottoms, and therefore anything represented 
as resting in them would have to be held in place in some incon­
spicuous manner, as the slender metal rods do hold the large amber 
pieces. But why amber in baskets? The amber, as we now know with 
certainty, is an import from the north; amber is not only sacred to 
Apollo, the Celts held it to be the tears which Apollo wept when he 
was banished to the land of the Hyperboreans.41 That amber was 
among the gifts sent by the Hyperboreans to Delian Apollo was first 
suggested by Welcker,42 and after reviewing the evidence in some 
detail, Cook43 has concluded not only that both the Western44 and 
the Eastern45 routes of the Hyperboreans coindde with established 
amber routes, but that "the cult of Apollon carne into Greece along 
the older (i.e. Western) Amber Route from the lands of the Hyper­
boreans, which is best located near the source of the Danube." 

If amber was among the Hyperborean gifts, then it was wrapped in 
wheaten straw,46 and if the Tiryns ornaments are cult representations 
of these sacred gifts to Apollo, then the gold wire baskets will be 
imitations of straw baskets of Celtic design. And the amber pieces 
which seem to resemble bones? Who would not think of the "famous 
hecatombs of asses" which Perseus found the Hyperboreans sacrifidng 
to Apollo?47 

Seltman's argument 48 that the Hyperboreans were not Barbarians 
but early Greek colonists in the upper Danube region is not relevant 
to Welcker's hypothesis concerning the congruence of the route of the 

39 Philadelpheus, supra n.7. 
40 Karo, supra n.2. 
n Apollonius Rhodius, Argonaunca 4.609--610. 
U F. G. Welcker, Griechische Gotterlehre II (Gottingen 1860) 353ff. 
43 A. B. Cook, Zeus II (Cambridge 1925) 497-499 and plate XXVI. 

u Hdt. 4.33. 
46 Paus. 1.31.2. 
46 Hdt., supra n.44. 
47 Pindar, Pyth. 10.32-34. 
48 C. T. Seltman, CQ 22 (1928) 155-159. 
2-G.R.B.S. 



18 ANALYSIS AND PROVENIENCE OF AMBER, II 

Hyperborean offerings with the amber routes, and his claim that 
Hthis hypothesis receives no support from the exhaustive researches 
of J. M. de Navarro" is unfounded. First, as far as the route described 
by Herodotus can be fixed geographically, it is as compatible with the 
Hwestern" amber route of Cook 49 as it is with the H eastern" amber 
route of de Navarro,5o since both approach the head of the Adriatic 
from a northeasterly direction. Second, de Navarro does not deal with 
later amber routes which may have connected the Vistula with the 
Dniester to the Black Sea; hence he has nothing to say about the 
reality of the Pausanian route for the Hyperborean offerings. 

The fact that the Tiryns crowns can now be dated, from the shape 
of the amber beads, close to 1100 B.C. adds further support to the 
congruence of the Hyperborean route with de Navarro's Adriatic 
amber route which opened in the Early Iron Age. De Navarro's con­
tribution to our knowledge of this route lies not only in having fixed 
its terminus post quem but in having shown that it emanated from the 
Eastern Baltic rather than from Jutland as had been thought before. 

At the same time it is now clear that this route cannot have been 
responsible for the earlier Mycenaean amber artifacts which will form 
the subject of subsequent reports in this series. 

Our interpretation will quickly find its critics. They must take into 
account another bit of evidence which points to the north: Marinatos51 

has pointed out the close similarity of the plaiting technique of the 
Tiryns objects with the Height-shaped" gold wire ornaments which 
were found at Cernilov near Hradec Kralove in Bohemia and which 
are now in the National Museum in Prague.52 The Czech ornaments 
belong to the Lusatian culture, which is assigned to the transition 
from the second to the first millennium B.C. Neither a stylistic relation­
ship to Celtic basketry nor to Lusatian jewelry need imply that the 
Tiryns ornaments were made outside Greece. That might best be 
decided by a technical study of the gold itself, for which Marinatos53 

has rightly called. We have been fortunate enough to obtain an anal­
ysis of the gold of the Tiryns objects from Dr Axel Hartmann of the 
Arbeitsgemeinschaft fUr Metallurgie am Wiirttembergischen Landes­
museum, Stuttgart, who is engaged in a comprehensive study of the 

&II A. B. Cook, supra n.43. 
60 J. M. de Navarro, GeogrJ 66 (1925) 481-507. 
n Marinatos, supra n.36. 
61 Anon., The Prehistory of C{echoslovakia (Prague 1958) plate XVll. 

58 Marinatos, supra n.36. 



C. VV. BECK, G.C. SOUTHARD, A. B. ADAMS 19 

impurity patterns of archaeological gold. By kind permission of Dr 
Hartmann,54 we append his analysis of the Tiryns gold in Table 3. 

TABLE 3. ANALYSIS OF THE GOLD VVIRE FROM THE TmYNs GOLD-AND-AMBER 

OB)ECTS55 

Element 
Silver 
Copper 
Platinum 
Lead 
Tin 
Nickel 
Bismuth 

Per cent 
11.0 
2.9 

.043 
less than 0.03 

0.024 
less than 0.01 
less than 0.01 

The German group has not yet analyzed a sufficient number of 
gold objects from the Lusatian or from the Mediterranean areas to 
place the Tiryns gold firmly. Dr Hartmann points to the high plati­
num content as an unusual feature which may permit assignment of 
the provenience of this gold when enough comparative material is 
available. Thus there is hope that another aspect of archaeological 
chemistry will soon contribute to the solution of the enigma of these 
unique objects. 

VASSAR COLLEGE 

December, 1967 

54 A. Hartmann, by private communication. 
55 The silver content is given in weight per cent of the alloy; all other values are weight 

per cent relative to gold = 100. 

ACKNOWLEDGEMENTS 

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 of the American 
Philosophical Society (Johnson Fund). 

The collection of samples in Athens was made possible by the assistance and co-operation 
of Professor H. S. Robinson of the American School in Athens, Dr J. Kontis of the Greek 
Archaeological Service and Dr V. Kallipolitis, Director of the National Museum. We are 
grateful for permission to reproduce the Athens National Museum photograph no. 3490/1. 

We thank Professor Christine Mitchell Havelock of Vassar College for generous advice 
and friendly criticism throughout this study, and Professor Emily Vermeule of Wellesley 
College for reading the manuscript.