Analysis and Provenience of Minoan and Mycenaean Amber, I Beck, Curt W Greek, Roman and Byzantine Studies; Fall 1966; 7, 3; ProQuest pg. 191

Analysis and Provenience of Minoan 
and Mycenaean Amber, I 

Curt W. Beck 

ECHAEOLOGY and the physical sciences can hardly be said to be 
properly married even now, but their unofficial union is of 
long standing, and it has been a fruitful one even without 

benefit of academic clergy. Yet one must admit that some of their 
offspring are illegitimate in more than one sense of the word. Among 
these are the several attempts to determine the provenience of amber 
artifacts by physico-chemical means. 

The geographic origin of archaeological amber finds in southern 
Europe is of some interest to the study of trade and cultural relations 
in early times, and nowhere more so than in Greece, where impres­
sive amounts of amber ornaments have been found in Mycenaean 
chamber tombs, shaft graves, and tholoi. When the first of these 
were found, the once unquestioned inference that amber from digs 
in southern Europe ipso facto meant importation from the distant 
North was no longer secure. The discovery of more or less substantial 
deposits of amber-like fossil resins in Bohemia, Austria, Hungary, 
Rumania, Italy, France, Spain, Portugal, and the Levant had raised 
the inevitable question whether some of the older amber artifacts 
found in southern Europe might not have been fashioned from such 
local or nearby deposits rather than from the celebrated Gold of the 
North. That question was first asked by Giovanni Capellini at the 
Seventh International Congress of Anthropology and Prehistoric 
Archaeology in Stockholm in 1874.1 Capellini's argument centered on 
the possible use of Bolognese amber in the early north Italian cultures 
of Villanova and Marzobotto. It found little response among the 
members of the congress. Only Cazalis de Fondouce pleaded for an 
open-minded consideration of Capellini's hypothesis and pointed out 

1 G. Capellini, Congres International d'Anthropologie et d'Archeologie Prehistoriques. 
Compte rendu de la 76 Session, Stockholm, 1874 (Stockholm 1876) 789-792, 799-800, 
803-809. 

191 



192 ANALYSIS AND PROVENIENCE OF AMBER, I 

that the stone-age amber artifacts in France might likewise have been 
made from the fossil resins native to that country.2 

Yet, while the dust quickly settled in anthropological and archaeo­
logical circles, it catalyzed a long series of chemical investigations of 
fossil resins by Otto Helm, an apothecary in Danzig. Those who 
cherish noble notions of the objectivity of science will be saddened to 
find that Helm did not take up his work with an open mind: he set 
out explicitly3 to disprove Capellini's contention, which he misunder­
stood or misrepresented as dealing with the origin of amber in 
Etruscan tombs. Helm's writings are colored throughout by a strong 
conviction that Baltic amber and only Baltic amber was the raw 
material of prehistoric and early historic artifacts in every part of the 
ancient world. In the next ten years, Helm published as many 
papers dealing with the properties of Baltic and southern amber and 
with the provenience of Italian amber artifacts.' During this period 
he developed the methods he then applied to Greek amber finds. 

Amber from Mycenae 
In the spring of 1883 Helm visited the National Museum in Athens 

and saw the amber beads which Schliemann had found in the first, 
third, and fourth shaft graves of the Late Helladic I Grave Circle A in 
the citadel of Mycenae.5 He applied for permission to analyze them. 
His first sample consisted of two grams of various fragments and gave 
inconclusive results. He then obtained a second sample which was a 
single large fragment weighing 4.1 grams. Unfortunately, the exact 
findspots are not indicated for any of these samples. Helm's findings 
are reported in two letters dated November 20 and December 17, 
1884.6 He describes the color of his pieces as "dark hyacinthine red," 

2 Cazalis de Fondouce, ibid. 809-810. 
8 O. Helm, "Notizen fiber die chemische und physikalische Beschaffenheit des Bern­

steins," ArchPharm 211 (1877) 229-246. 
~ E.g. (a) O. Helm, "Uber sicilianischen und rumanischen Bernstein," &hriften der Natur­

forschenden Gesellschaft zu Danzig, N.F. 5 nos. 1-2 (1881) 293-296; (b) idem, "Uber sicilianischen 
Bernstein," ibid. 5 no. 3 (1882) 8-9; (c) idem, "Uber die element are Zusammensetzung des 
Ostsee-Bemsteins," ibid. 5 no. 3 (1882) 9-11; (d) idem, "Uber Appeninnen-Bemstein," 
ibid. 5 no. 3 (1882) 11-14. 

5 H. Schliemann, Mycenae (New York 1878) 203-204, 245,308; German ed. (Leipzig 1878) 
235, 283, 353. 

8 O. Helm in H. Schliemann, Tiryns (New York 1885) 369-372; O. Helm, "Uber die 
Herkunft des in den alten KOnigs-Grabern von Mykenae gefundenen Bemsteins und fiber 
den Bemsteinsauregehalt verschiedener fossiler Harze," Schriften der Narurforschenden 
Gesellschaft zu Danzig, N.F. 6 no. 2 (1885) 234-239. 



CURT W. BECK 193 

"wax-colored," "bright ruby-red," "bright orange-red," and "almost 
as clear as water," 7 and considers these in keeping with Baltic origin. 
He cites their hardness (not given numerically), their odor on burning, 
and the appearance of their weathering crust as being like those of 
Baltic amber. His main piece of evidence is the semi-quantitative 
determination of succinic acid. Helm had found this acid present in 
authentic samples of Baltic amber in amounts ranging from 3.2 % 
to 8.2 %,8 but absent in most, though not all, authentic samples of 
non-Baltic amber.9 His first Mycenaean sample yielded only 1.6 %. 
Helm explained this, as well as the unusually high ash content of 
3.2 %, by the extensive weathering which had led to "loss and 
change" of organic constituents and to accretion of mineral matter 
from the surrounding soil. While this sounds plausible enough, it is 
in direct contradiction of his own earlier findings that the succinic 
add content of Baltic amber is about twice as great in the weathering 
crust (8.2 %) as it is in unweathered clear-yellow amber (3.2-4.5 %).10 
It was this low succinic acid assay which moved him to ask for the 
second, better preserved samples in which he found 6.0 % succinic 
acid. Finally, Helm conducted an elemental analysis of the second 
sample; the results closely agreed with earlier analyses of Baltic 
amber: 

TABLE 1. HELM'S ELEMENTAL ANALYSIS, BY PER CENT 

Element Mycenae Bead Baltic Amber 

Carbon 7S.60 7S.63 

Hydrogen 10.0S 10.48 

Oxygen (by difference) 10.98 10.47 

Sulfur 0.34 0.42 

On this evidence, Helm had "no hesitation declaring the amber 
beads found in the royal tombs of Mycenae to be Baltic amber; there 
are no facts known to show that any product corresponding to the 
above results can be found elsewhere, i.e. anything chemically and 
physically resembling amber." Helm was quite likely right in his 
assignment. The number and size of the beads in the Grave Circle A 

7 German "fast wasserklar"; since Helm refers back to this description explicitly as a 
color, he must mean 'almost colorless' rather than 'almost transparent'. 

8 Helm, supra n.3. 
9 Helm, supra n.4. 
10 Helm, supra n.3. 



194 ANALYSIS AND PROVENIENCE OF AMBER, I 

tombs are in themselves arguments for their Baltic origin. His tests 
heighten that probability to a degree where it may well seem churlish 
to quarrel with it. Yet there remains the disturbing fact that Helm 
presented his conclusions as following inexorably from his chemical 
tests. They do not. 

Helm's main piece of evidence is here, as in other examples of his 
work, the semi-quantitative determination of succinic acid. It is an 
objective test which can answer specific questions decisively. For 
example, Helm had earlier analyzed four samples of amber native 
to the province of Emilia and seven amber artifacts from the same 
region which he had obtained from the Museum of Bologna.ll The 
total absence of succinic acid in the four geological samples and the 
invariable presence of it in the archaeological samples (4.8-6.3 %) 
support the conclusion that the seven artifacts were not fashioned 
from local raw material, although the modern statistically-minded 
archaeometrist should like to see a much larger number of analyses 
of geological samples before admitting to the universal statement 
that Hall native Apennine amber is devoid of succinic acid." But the 
evidence does not show that no north Italic artifacts were fashioned 
from local Bolognese amber; nor does it show that the seven analyzed 
artifacts were made of Baltic amber. Yet Helm was much inclined 
to draw both of these conclusions, and in doing so he raised expecta­
tions about the nature of the contribution which science can make to 
archaeology which quite exceed the possible. 

Helm persistently overrated the significance of a single experiment, 
while caution would warn that it is at best one small piece of a very 
large jigsaw puzzle. There are thousands of amber artifacts in the 
Mediterranean countries. They cover a large area and a long period of 
time. A sizable and representative fraction of them must be analyzed 
before it is reasonably safe to settle the kind of question which Capel­
lini has asked and which Helm answered all too hastily. 

Helm's second error was to declare an artifact to be of Baltic origin 
if it contained sufficient succinic acid. That would be valid if it were a 
fact that only Baltic amber contains comparable amounts of succinic 
acid. Now this is not a fact, and Helm well knew it. He had himself 
found 5.2 % succinic acid in a sample of Rumanian amber,12 but he 
nevertheless ruled out a non-Baltic origin for the Mycenae beads on 

11 Helm, supra n.4 Cd). 
12 Helm, supra n.4 (a). 



CURT W. BECK 195 

the grounds that "Rumanian amber ... is easily distinguished by a 
competent judge from Baltic amber by color, hardness, and disin­
tegrated layer." 

This is fairly characteristic of all the early work on amber: objective 
measurements are given great weight when they appear to be use­
ful, but they are quickly retired in favor of the subjective dedsion of 
a "competent judge," who is invariably the author himself, when 
they raise problems or contradictions. The fact is that the succinic 
add analysis only permits assignment of an amber artifact to one of 
two groups: those which are known to contain the acid (including 
Baltic amber, some Sicilian amber, some Rumanian amber, and 
several varieties from Portugal, France and Galicia 13), and those which 
are known to lack it. Even this dichotomy is imperfect since there 
is a sizable group of fossil resins which have never been tested for 
succinic acid.14 

Similar limitations apply to the identification by elemental compo­
sition. There is a considerable literature of quantitative analyses for 
carbon, hydrogen, oxygen (by difference), and sometimes sulfur in 
fossil resins. In view of the heterogeneity of resins, and of the changes 
which they have suffered during millions of years of storage under 
diverse geological conditions, it is not surprising to find that the com­
position varies rather widely within each group of samples, and that 
agreement of such values gives a clue, but never a proof, of identity. 
The same must be said of color, specific gravity, melting point, re­
fractive index, fluorescence, and every other property which has ever 
been used to distinguish fossil resins from one another. 

Thus the close consideration of Helm's pioneering work allows us to 
assess the methodological problems of any physico-chemical approach 
to the provenience of amber artifacts. Let us assume that all naturally 
occurring fossil resins have been tested for five properties, e.g. 
succinic acid content, elemental composition, color, hardness, and 
odor on burning. These may not be the most promising properties 
to choose, but they are the ones which Helm, in fact, chose. 

(a) If no species of fossil resin agrees with any other in all five 
properties, then five tests will identify an artifact of unknown origin 
decisively. 

13 w. LaBaume, "Zur Naturkunde und Kulturgeschichte des Bernsteins," Schriften der 
Naturforschenden Gesellschaft ZU Danzig, N.F. 20 no.1 (1935) 5-48. 

14 M. H. Hey, Chemical Index of Minerals (London 1962) sec. 33. 



196 ANALYSIS AND PROVENIENCE OF AMBER, I 

(b) If some species have all five properties in common, then five 
tests will rule out some places of origin for an artifact, but no positive 
identification can be made. 

Life is never as neat as logic. First of all, our primary assumption 
has not yet been met. Many varieties of fossil resins have not been 
characterized.1s Secondly, the properties which Helm and others have 
used are not points on a scale; they are ranges, and overlapping ranges 
at that. There is thus no more than an a priori probability that a 
given sample would have a value within a particular range if it came 
from the geographical area which is characterized by that range. A 
more useful statistical concept is the null hypothesis, i.e. the chance 
of finding a value within the range even though the sample does not 
come from the area characterized by that range. If the null hypothesis 
is 0.5 (meaning that there is an even chance that the property might 
fall within the range even though the sample did not come from the 
area), then agreement in five properties would give an over-all null 
hypothesis of (0.5)5 = 0.031; hence the degree of confidence that the 
artifact does come from the area would be almost 97 %. This is a 
highly satisfactory degree of certainty, and its attainment is in prin­
ciple possible. The only final limitation is that there may be deposits 
of fossil resins which were worked in antiquity and which were either 
completely exhausted or which have not been rediscovered in modem 
times. An artifact from such a source would be in principle unidenti­
fiable, and it might easily be falsely assigned to a known deposit 
yielding resin of closely similar or even identical properties. This is 
too unlikely to discourage work in this field, but it should help us to 
maintain a modicum of modesty in advancing claims of positive 
identification. 

Additional tests on Schliemann's amber from Mycenae and Troy 
were reported by Rossler in 1899.16 He had inferred from the litera­
ture, and confirmed by experiment, that amber usually contains 
some sulfur while African copals never do. The qualitative test for 
sulfur in amber is quite simple: on heating in a test tube, the sulfur 
escapes as gaseous hydrogen sulfide which can be identified by its 
highly offensive odor or, better, by the black-brown discoloration 
which it imparts to a piece of paper soaked in a solution of lead 

15 Hey, supra n.14. 
16 O. Rossler, "Chemische Unterscheidung von Bernstein und Kopal," ArchPharm 237 

(1899) 239-240. 



CURT W. BECK 197 

acetate. All this had, indeed, been established by Baudrimont in 
1864,17 and later analyses showed that there is sulfur not only in 
Baltic amber, but also in Sicilian, French, Hungarian and Rumanian 
amber. Rossler's tests, therefore, prove only what no one had ever 
seen cause to doubt: that the artifacts are some kind of amber and 
not African copal. The tests throw no light on the geographical 
origin of the amber beads. 

Afnber from Kakovatos 
The second substantial experimental study of Mycenaean amber 

was made by Jonas18 in 1908 on a single large (15 x 42 mm) bead 
from a LM IB tho los near Kakovatos which had been excavated by the 
German Archaeological Institute.19 From Muller's list of finds,20 the 
bead must have come from tho los A. Like Helm, Jonas did all his 
work on one sample and was forced to assume that the origin of one 
bead reflects that of all the others. He conducted tests on the bead 
and, simultaneously, on three reference samples: 

(a) an average sample of Baltic amber selected from 500 grammes of 
"clear" material drawn from the Prussian State Amber Works at 
Konigsberg, 

(b) a single bead of Sicilian amber (simetite) from Catania, taken 
from the collection of the University of Konigsberg, and 

(c) a single small piece of Rumanian amber (rumanite), also from 
the university collection. 

Jonas eVidently expected that the physical and chemical properties 
of the archaeological sample would agree with those of one of the 
three geological samples. The fallacy of this unwarranted restriction 
(made again by Reutter in 1916 in his work on amber from Swiss 
lacustrine settlements 21) is obvious: the archaeological sample may be 

17 E. Baudrimont, "Du soufre considere comme l'un des elements CQnstituant du suc­
cin," CR HebdAcadSc Paris 58 (1864) 678-680; idem, BullSocChim France, SER. II 1 (1864) 328-330. 

18 R. Jonas, "Bernsteinperlen aus einem mykenischen Kuppelgrabe und die Identi­
fizierung ihrer Substanz mit Succinit," Schrifren der Physikalisch-okonomischen Gesellschaft 
zu Konigsberg 49 (1908) 351-368. 

19 (a) W. Dorpfeld, "Tiryns, Olympia, Pylos," AthlVIitt 32 (1907) ix; (b) idem, "Alt·Pylos, 
I. Die Kuppelgraber von Kakovatos," ibid. 33 (1908) 295-317; (c) K. MUller, HAlt-Pylos,II. 
Die Funde aus den Kuppelgrabern von Kakovatos," ibid. 34 (1909) 269-328. 

20 K. MUller, supra n.19 (c). 
21 L. Reutter, "Des ambres lacustres, "CR HebdAcadSc Paris 162 (1916) 421-423; D. Viollier 

and L. Reutter, "Contribution a l'etude de l'ambre prehistorique," AnzSchweiz N.F. 18 
(1916) 169-182. 



198 ANALYSIS AND PROVENIENCE OF AMBER, I 

assigned to one of the tested varieties with which it most nearly agrees, 
while it may actually belong to one of the many varieties which were 
not considered for comparison. 

Jonas rejected Helm's method of analysis for succinic acid (pyrolysis 
of the resin and collection, purification, and weighing of the succinic 
acid deposited in the neck of the retort) on two grounds: first, because 
he had insufficient material and, secondly, because he felt that "it 
gave no guarantee that the isolated succinic acid was an actual 
constituent of the amber and had not been formed in the course of 
pyrolysis as a fragmentation product of higher-molecular acids." 
There can be no quarrel with the first point, but the second one is 
specious: whether succinic acid pre-exists in amber or is formed in the 
course of destructive distillation (pyrolysis) is a very real question for 
the chemist who probes the molecular structure of resins, but it is 
irrelevant to the problem of provenance. All that matters here is 
that some ambers do furnish succinic acid on heating while others 
do not. 

Jonas states that the Kakovatos bead had the same odor on rubbing 
and heating as Baltic amber. Neither the bead nor his Baltic reference 
sample would melt at temperatures up to 360°C, although (as Jonas 
knew) Helm had reported 22 the "melting points" of Baltic amber to 
lie between 287° and 300°C. Fossil resins generally do not melt; they 
decompose, and decomposition points are notoriously difficult to 
reproduce. Even so, a difference of nearly 80°C is curious. 

Jonas next turned to a comparison of solubilities in 96 % ethanol. 
He made single determinations of the Kakovatos sample (18.58 % 
soluble), his Baltic reference sample (21.91 %), and the Sicilian sample 
(14.41 %). His Rumanian sample was too small for a solubility test, 
and he cites earlier work by Murgoci23 which gives the solubility of 
rumanite in 96 % ethanol as 6 %. These tests were thus inconclusive, 
which comes as no surprise. The solubility of Baltic amber in ethanol 
has been given as 10-12 % by Heyer,24 17-25 % by Helm,25 and 14.3 % 
by Coffignier.26 The solubilities of other fossil resins vary similarly, 
demonstrating once more that the wide and overlapping ranges of 

22 Helm, supra n.3. 
23 G. Murgoci, "Gisements du succin de Roumania ... ," Memoriile Congresului de laJasi 

(Bucharest 1903) 1-34. 
Z4 J. c. H. Heyer, Chemische Versuche mit Bernstein (Erfurt 1787). 
25 Helm, supra n.3. 
26 G. Coffignier, "Sur la solubilite du succin," Bul/SocChim France SER.IV 5 (1909) 1101-1103. 



CURT W. BECK 199 

single physical properties are inadequate for the characterization of 
these complex and heterogeneous mixtures. 

In the end, Jonas' identification of the Kakovatos bead with Baltic 
amber (mineralogically: succinite) rests on his experiments with 
quantitative cleavage (chemically: saponification) by means of sodium 
methoxide, CHaONa, and back-titrating the excess base with standard 
solutions of dilute sulfuric acid. His results are given somewhat 
unconventionally in terms of grams of sulfuric acid equivalent to the 
amount of sodium methoxide which are consumed by one gram of 
resin. They are easily converted to the commonly cited so-called 
saponification numbers, defined as the number of milligrams of 
potassium hydroxide (even though another base, such as sodium 
methoxide, may have been used in the actual test) which are con­
sumed by one gram of resin. 

The saponification number (S) is the sum of two other quantities: 
the acid number (A), defined as the number of milligrams of potas­
sium hydroxide consumed by one gram of resin at room temperature 
by the neutralization of free acids in the sample, and the ester number 
(E), defined as the number of grams of potassium hydroxide required 
to cleave all ester linkages in one gram of resin, a reaction which 
requires heating to be completed in a reasonable length of time. Thus 
S = A + E. Of these three quantities, S and A are determined directly 
by experiment, E is calculated from them. The relative values of these 
numbers have been of use in classifying fossil resins; most recently 
Chauffin 27 has published some careful work on the acid number of 
French fossil resins. Jonas determined only saponification number: 

TABLE 2. JONAS' ANALYSIS BY QUANTITATIVE CLEAVAGE 

Sample 

Kakovatos bead 
Baltic amber (succinite) 
Sicilian amber (simetite) 
Rumanian amber (rumanite) 

Sulfuric acid eqUivalent Saponi-
to the base consumed fication 
by one gram of resin Number (S) 

0.l323 g. 151.4 
0.1285 g. 147.0 

0.0945 g. 108.1 

0.1273 g. 145.6 

Jonas concedes that these results fail to distinguish between Baltic 
and Rumanian amber, and that the Kakovatos bead might hence be 

27 J. Chauffin, BullMenSocLinneenne Lyon 31 (1962) 189-191 and 32 (1963) 103-105. 



200 ANALYSIS AND PROVENIENCE OF AMBER, I 

identical with either. He nevertheless identifies it with Baltic amber 
on the grounds that both changed in the course of saponification to a 
"spongy, voluminous mass which shows all the properties of the 
alcohol-alkali salt of succinoresinol described by Aweng." Jonas 
does not state what these properties are. Aweng's28 succinoresinol 
is a very vaguely characterized fraction derived from the saponi­
fication of the alcohol-insoluble portion of succinite. Jonas goes on to 
make an elaborate calculation of the amount of succinic acid which he 
claims his tests show to be present: allowing properly for the fact 
that his supposed succinoresinol was derived from whole amber while 
Aweng's was obtained only from the alcohol-insoluble portion (first 
called 'succinin' by Berzelius 29), he calculates the amount of base 
required to saponify the alcohol-insoluble portion of his samples. 
Then, assuming that his resins have been cleaved into equivalent 
amounts of succinoresinol and succinic acid, he calculates the amount 
of succinic acid which appears to be present in his sample of succinite 
to 6.6 %, and in the Kakovatos bead to 7.0 %. 

This adventurous extrapolation is quite untenable. Jonas simply 
begs the question. The calculation of 7 % succinic acid in the Kako­
vatos bead assumes not only that succinic acid is present, but that 
succinoresinol is present as well; in short, that the Kakovatos bead is 
succinite. At no time did Jonas prove the presence of succinic acid in 
any of his samples. He does claim a qualitative test for "not incon­
siderable amounts" of succinic acid by the formation of the iron salt 
with ferric chloride from the pyrolysates and hydrolysates of all his 
samples save simetite. No such test is recognized. Succinic acid does 
form an insoluble salt with ferric ions, but so do virtually all organic 
acids. A simple reductio ad absurdum will show the unreasonableness 
of Jonas' procedure: application of his mathematical operations to his 
sample of simetite will lead to the startling conclusion that it contained 
4.9 % succinic acid, and hence could not have been simetite. 

Jonas would reply to this argument that his sample of simetite 
did not, after all, produce a "spongy, voluminous mass" on saponi­
fication while both succinite and the Kakovatos bead did, and that the 
latter two reacted with ferric ion while the simetite did not. That is 
true, but it must then be said that Jonas' identification of the Kako-

18 E. Aweng, ObeT den Succinit (Diss. Bern 1895) and ArchPharm 232 (1894) 660-688. 
al J. J. Berzelius, "Einige Bemerkungen tiber den Bernstein," VetenskAcadHandling 1827, 

and Poggendorffs Annalen der PhYSik, 2.p 12 (1828) 419-429. 



CURT W. BECK 201 

vatos bead with Baltic succinite rests solely on the appearance of the 
"spongy, voluminous mass," since even the dubious ferric chloride 
test lumps the Kakovatos bead together with both rumanite and suc­
cinite, as do the numerical results of his saponifications. 

As with Helm's work on the beads from Grave Circle A at Mycenae, 
Jonas' tests substantially increase the a priori probability that the 
analyzed materials are of Baltic origin, but they do not, as they 
claim to do, furnish any conclusive, objective, "scientific" proof to 
that effect. 

There are several other Mycenaean amber finds which have 
apparently not been analyzed. The most remarkable among them are 
the curious gold-and-amber "diadems" from Tiryns.3o Lesser pieces 
are known from Nauplia,31 Menidi (Attica),32 and Ialysos on Rhodos.33 

Wace found large quantities in the chamber tombs of Mycenae.34 

Amber from Crete 
The large amounts of amber found at some Mycenaean sites con­

trasts sharply with its pauoty in Minoan Crete. Of particular interest 
are "two small bits of amber" found in a primitive tholos at Porti 
in 1906.35 They were subject of a disappointingly superficial exami­
nation by Mosso,36 who simply burned one of them in an open flame. 
On the slight evidence that it melted and burnt with a pleasant odor 
without leaving any ash, Mosso declared it to be Baltic amber. Evans37 

very properly criticized Mosso's attribution by pointing out that the 
test made "falls very far short of a scientific analysis, and the behavior 
of the burnt fragment was only such as might have repeated itself in 
the case of a piece of simple resin," by which he means a non-fossil 

30 A. Philadelpheus, ArchDelt 2 (1917) App. p.17. 
311. Kondakis, Athenaion 8 (1879) 521-523; D. Philios, ibid. 523-526. 
32 A. Furtwangler and George Loeschcke, Mykenische Vasen (Berlin 1886) 40. 
33 Supra n.32, pI. B 12 and 13. 
34 A. J. B. Wace, Chamber Tombs at Mycenae [Archaeologia 82] (Oxford and London 1932) 

86, 204-205, plates ix, xxix, xxxvi. 
35 S. Xanthoudides, The Vaulted Tombs of Mesard, trans!' J. P. Droop (London 1924) 69. 
38 A. Mosso, Le Origini della civilta mediterranea (Milan 1910) 291-292; English transi. 

M. C. Harrison, The Dawn of Mediterranean Civili~ation (New York n.d.) 368-369. Mosso here 
erroneously reports the pieces as corning from Kumasa instead of Porti, and Evans follows 
him infra n.37 (a) but not n.37 (b). 

37 (a) A. Evans, "The Tomb of the Double Axes and Associated Group," Archaeologia 
65 (1914) 44; (b) A. Evans, in preface to Xanthoudides (supra n.35) p. xii. 



202 ANALYSIS AND PROVENIENCE OF AMBER, I 

resin of local origin such as the Minoans apparently used as a sort of 
incense. If these fragments were indeed Baltic amber, they would 
represent the oldest example of it in the eastern Mediterranean. 
Moss038 dated the tholos to the end of Early Minoan or the begin­
ning of Middle Minoan; Evans39 chose EM TIl, or about 400 years 
before the next well-attested occurrence of amber in Greece. Un­
fortunately, there can be no re-examination, since the pieces seem to 
have been destroyed.40 

Evans himself found an amber disk edged with gold and two per­
forated beads in a chamber tomb at Is6pata.41 He sent part of one 
bead to Otto Olshausen in Berlin, who had taken up the analysis of 
amber artifacts and continued it after Helm's death in 1902. Olshausen 
declared it to be Baltic amber, but since he did not publish details of 
his experimental work, it is difficult to assess the evidence. Evans 
reports the conclusion as having been based on (a) the appearance of 
the weathered bead, (b) the formation of hydrogen sulfide during dry 
distillation (pyrolysis), and (c) the acidic reaction of the distillate. He 
adds that a more complete analysis could not be made because of 
"some defect in the coal employed." It must be said that a great many 
fossil resins other than Baltic amber contain sulfur42 which will be 
converted to hydrogen sulfide on heating, and that an acidic reaction 
of the distillate may be caused by any volatile acid, e.g. the formic 
acid John found in Lebanese schraufite.43 However, some of Ols­
hausen's work notes have been posthumously published by 
LaBaume,44 and from them it is clear that the evidence was somewhat 
stronger than Evans reported: Olshausen observed "masses" of 
crystals in the neck of the distillation vessel. This is a strong indication 
for the presence of succinic acid, but since some mishap prevented him 
from identifying these crystals pOSitively, he did not include them in 
his report. Evans made a bare reference to a single amber bead from 
an LM I or TI find at Arvi in southeastern Crete, but its whereabouts is 
not known. 

38 Mosso, supra n.36. 
39 Evans, supra n.37 (a), 44. 
40 Evans, ibid. 
41 Evans, supra n.37 (a), 42-43. 
42 Hey, supra n.14. 
t3 K. John, "Bernstein und Schraufit aus dem Libanon," VerhandlKKGeolReichsanstalt 

1876. pp.255-257. 
U LaBaume, supra n.13. 



CURT W. BECK 203 

Post-Mycenaean Amber 
It is well known that little amber has been found in the post­

Mycenaean Greek world. Isolated finds have been made at Sparta45 
and Eleusis46; larger quantities come from archaic Ephesus.47 Later 
and far afield are two bits of amber from Tumulus II at Gordian, 
dated to the early sixth century B.C.48 The smaller of these was sent 
to Helm by way of Olshausen and the results reported to the Kartes 
by the same route: the weathering crust of the sample contained 3.3 % 
succinic acid. Again Olshausen's notes49 give additional details from 
two letters written to him by Helm: the material showed a remark­
ably high <melting point' of 386°C. It is characteristic that even before 
he had carried out the quantitative analysis, Helm declared the sample 
to be succinite in his first letter on the basis of its appearance and its 
odor on burning. 

Analysis by Infrared Spectroscopy 
Our review of past work on the chemical identification of Greek 

amber artifacts shows that only a handful of samples has been 
analyzed, and that they were analyzed between 1884 and 1908 by 
methods which leave generous room for doubt.50 That the conclusions 
have been generally accepted as reliable by archaeologists is not the 
archaeologists' fault; quite clearly, the analysts announced their 
findings with the authoritative assurance which is always expected of 
them, but which is not nearly as often justified. Only Professor 
Marinatos has recently expressed some healthy scepticism 51 and said 

45 R. M. Dawkins, "Laconia, I. Excavations at Sparta, 1907. The Sanctuary of Artemis 
Orthia," BSA 13 (1906/07) 73-75, 83-84 and pIA; J. P. Droop, ibid. 112-113. 

46 ArchEph 1885, p1.9,3, cited by G. Karo, "Die Schachtgraber von Mykenai," AthMitt 40 
(1915) 157 n.l. 

47 D. G. Hogarth, Excavations at Ephesus (London 1908) 213-216, plates 47 and 48. 
'8 G. and A. Korte, Gordion: Ergebnisse der Ausgrabung imJahre 1900[JdI Suppl. 5] (Berlin 

1904) 128-129. 
49 In LaBaume, supra n.13. 
50 O. Olshausen, "Uber den alten Bernsteinhandel und die Goldfunde, II," ZfE. Verh 

1891, pp.286-319, cites evidence which leads him to conclude that "the method of analysis 
has uncertainties which make a comparison between the results of different analysts 
difficult" and that "dry distillation (pyrolysis) is not well suited for analytical procedures." 
According to the data given in this reference, Helm conSistently found more succinic acid 
than other analysts; this may be a credit to his skill, it is none to the method. 

61 S. Marinatos, "The Minoan and Mycenaean Civilization and its Influence on the 
Mediterranean and on Europe," Atti del VI Congresso Internationale delle Scienze Preistoriche 
e Protostoriche I (Rome n.d.) 166. 

2-G.R.B.S. 



204 ANALYSIS AND PROVENIENCE OF AMBER, I 

that "a new and comprehensive study of the amber question is 
needed, with a chemical examination of as many samples as possible, 
and coming from all areas available." 

Not the least of the shortcomings of the succinic acid assay as devel­
oped by Helm and Olshausen are the quantities of material (in the 
order of 1-2 grams) which are completely destroyed in the process, 
and the time and skill required for a single determination. They are 
without doubt the main reasons why the method has not been applied 
as widely as Marinatos rightfully demands.52 

We therefore set out, some years ago, to see if modern organic 
analysis might not offer ways of distinguishing between mineralogical 
varieties of amber which would require so little material, skill, and 
time that they could be applied to the hundreds if not thousands of 
amber artifacts which await identification. We have found such a 
method in infrared absorption spectrophotometry. Preliminary 
results 53 and technical details 54 have been published elsewhere. 
We will here give only a bare outline of the principles involved. 
A sample of about two milligrams is ground with a hundredfold 
excess of potassium bromide (a salt which is transparent to infrared 
radiation), and the mixture is pressed into a clear pellet in an evacu­
able die. The spectrum of the pellet is then determined in a com­
mercial infrared spectrophotometer. We have successfully used one 
of the simpler and less expensive instruments on the market, the 
Perkin-Elmer Model 237. The instrument records the absorption of 
infrared light as a function of wavelength over the range from 2.5 

to 16.0 microns (4000 to 600 em-I). The absorption curve is a source 
of much chemical information which need not be considered here55 ; 

for purposes of identification it is simply a sort of fingerprint of the 
sample. For pure compounds, identity of spectra establishes identity 
of chemical composition. For such complex mixtures of compounds as 
fossil resins, essential likeness of spectra is as much as can be expected. 
By recording the infrared spectra of a large number of fossil resins 
from every known deposit, we have found that Baltic amber has a 

&I J. M. de Navarro, "Prehistoric Routes between Northern Europe and Italy defined by 
the Amber Trade," Geogr] 66 (1925) 481-507, made the same demand for amber analyses 
on a large scale for the sake of general European prehistory. 

liS C. W. Beck, E. Wilbur and S. Meret, "Infrared Spectra and the Origin of Amber," 
Nature 201 (1964) 256--257. 

II C. W. Beck, E. Wilbur, S. Meret, D. Kossove and K. Kermani, "The Infrared Spectra of 
Amber and the Identification of Baltic Amber," Archaeometry 8 (1965) 96--109. 

66 See Beck et aI., supra n.54. 



CURT W. BECK 205 

distinctive absorption pattern in the region between 8 and 9 microns. 
In a well-preserved sample, the pattern shows a perfectly flat 
<shoulder' between s.o and S.5 microns, followed by a peak which 
centers upon 8.7 microns, after which absorption diminishes sharply 
(figure 1, Spectrum 74). In many cases, the normally horizontal 
shoulder has acquired a negative slope (figure 1, Spectrum 179). This 

74 

8 9 10 1 1 8 9 10 1 1 

(a) Baltic Coast (b) Heligoland 

Figure 1. INFRARED SPECTRA OF BALTIC AMBER FROM 7.5 TO 12.0 MICRONS 

change is invariably connected with a decrease in the intensity of an 
absorption peak at 11.3 microns, and we have traced both to the oxi­
dative degradation caused by long exposure to air. 56 It is therefore not 
surprising that many archaeological samples show spectra (figure 2) 
in which these changes have occurred to varying degrees, and that the 
changes are more pronounced in samples which have been exposed 
the longest. 

We think it important to state clearly the limitations of the present 
state of our knowledge. We hope, of course, to make further progress, 
but at this time it is only possible to say whether an amber artifact is 
Baltic or not. We have now recorded the spectra of nearly SOD amber 
samples of which about half are Baltic succinite and the other half 
are fossil resins from every part of the world. The latter show such 
variety in their spectra, and at the same time so much overlap 
among geographically separated groups, that much more evidence 

56 See Beck et a/., supra n.54. 



206 ANALYSIS AND PROVENIENCE OF AMBER, I 

must be gathered before a rational classification can be made. If the 
resins native to, say, Sicily and the Levant were formed by the 
same, or by closely related, species of trees, their chemical composi-

196 

8 9 10 11 12 8 9 10 11 12 

Megalithic Grave, ZiesendorJ Slavic Fortification, Rostock-
(Neolithic Age) Dierkow 

Museum der Stadt Rostock, Mecklenburg, Germany (DDR) 

Figure 2. INFRARED SPECTRA OF ARCHAEOLOGICAL SAMPLES OF BALTIC AMBER 

tion would be quite or nearly identical, and so would their spectra. 
Even when it is possible, classification will not necessarily solve 
burning archaeological questions. 

However, none of the spectra of non-Baltic European amber shows 
features which would allow any confusion with those of Baltic 
amber.57 We therefore feel very confident that an amber artifact 
which has the spectral features described above can be considered of 
Baltic origin. As mentioned before, there are Baltic samples which 
give such poor spectra as a result of oxidative degradation that a 
careful analyst should suspend judgement. Spectrum 194 in figure 2 
is an example. So far, these have been few, and more often than not 
repeated sampling of the intractable piece has eventually brought up 
a spectrum which made an unequivocal decision possible. On the 
basis of past experience, we have no hesitation in pronouncing a 

57 The qualification 'non-Baltic European' amber is necessary because we have found 
spectral patterns indistinguishable from those of Baltic amber in some North American 
fossil resins. Though that may be electrifying news to the palaeobotanist, it is a matter of 
indifference to the archaeologist. 



CURT W. BECK 207 

specimen to be Baltic if it gives only one characteristic Baltic spectrum 
among several inconclusive ones. 

The ease and speed of the spectroscopic technique now permits 
the analysis of every known amber artifact of archaeological interest; 
the minuteness of the sample required allows the analysis of even 
small and unique finds without destroying their scientific and aesthetic 
value. We have begun to undertake the spectroscopic analysis of all 
Minoan, Mycenaean, and Greek amber artifacts, and we will report 

8 9 10 11 8 9 10 11 

Figure 3. SPECTRA OF AMBER FRAGMENTS IN THE CHORA MUSEUM 

complete results in subsequent papers in this journal. At this time 
the only tests made are of a handful of samples from the Museum at 
ehora.liB The spectra are shown in figure 3. Sample no. 2335 comes 

58 C. W. Beck, "Determination of the Provenience of Amber by Infrared Spectrophoto­
metry," paper read at the 67th General Meeting of the AlA, Providence, R.I., 30 December 
1965. 



208 ANALYSIS AND PROVENIENCE OF AMBER 

from a shaft grave under Room 97 of the Palace of Nestor and is, of 
course, older than the palace. It is of Baltic origin. The other four are 
from the tholos V l' the Vayenas tholos, excavated by Lord William 
Taylour in 1957 and described in Professor Blegen's report for that 
year 59 and by Taylour.60 The latest pottery in this tomb belongs to 
the transition period LH II/III, but some is as old as MH. The samples 
numbered 2022A and 2046B are decidedly Baltic, but nos. 2043 and 
2055A are just as decidedly not. The spectra of the latter two are prac­
tically identical; it is more than likely that the two samples are 
fragments of the same artifact. They are the first non-Baltic amber 
found in a Mycenaean context. 

These few examples may indicate the modest questions which 
spectroscopic analysis can answer as well as those which remain. For 
the non-Baltic amber the extremely important question of origin is 
at this time beyond our knowledge. The spectrum leaves a number of 
possibilities open, among them some types of simetite from Sicily 
and schraufite from the Levant. Professor Marinatos has mentioned 
the possibility that some late Mycenaean amber may be of Sicilian 
origin 61; we hope to settle this by further experimental work. 

Amber Routes to Mycenaean Greece 
Another large question concerns the precise geographical origin of 

Baltic amber found in the Mediterranean region and the routes along 
which it traveled south. The very extensive literature on the amber 
trade stands in need of critical review. There is a great deal of literary 
and archaeological evidence for Roman times, but uncertainty con­
founded by speculation is all that can be offered for earlier periods. 
For one thing, the natural distribution of Baltic amber due to diluvial 
and alluvial geological events is more extensive than is often realized. 
It extends from England across northern Germany and southern 
Scandinavia to deep into Russia. The Phoenicians are plausibly 
credited with an early sea route from England or Jutland to the 
eastern Mediterranean.62 Professor Marinatos has suggested 63 that 
such trade may have reached far back into Minoan and Mycenaean 

58 C. W. Blegen, "The Palace of Nestor Excavations of 1957," AJA 62 (1958) 178. 
60 W. Taylour, The Mycenaeans (London 1964) 80-81. 
61 Marinatos, supra n.5!. 
6Z Homer, Odyssey 15.460. 
63 Marinatos. supra n.51. 



CURT W. BECK 209 

times and that the Aeolian Islands, particularly Lipari, were important 
way stations on this route. We have one small piece of evidence which 
fit his hypothesis: a large annular fragment from Capanna F on the 
Aeolian island of Salina was found in association with Hnecklaces of 
hard stone beads (carnelian) and of glass paste beads of Mycenaean 
manufacture imported from the Aegean ... and with pottery frag­
ments of the periods LH IIIA2 and IIIB (1400-1250 B.C.)."64 In the 
course of our nearly completed analyses of all Sicilian and Aeolian 
amber finds, we have established that this fragment is of Baltic 
origin.65 

Even so, an early eastern land route to Greece cannot be ruled out. 
The route need not have been very long, since Baltic amber has been 
carried to the very shores of the Black Sea at the end of the last Ice 
Age. It is worthwhile to bring the compilation of amber finds in 
Russia by Koppen 66 to the attention of archaeologists (figure 4). The 
map would surely have delighted Miss Bacon.67 It is true that the 
isolated finds at the mouths of the Dnieper can hardly have been the 
_source of the many large beads found at Mycenae and Kakovatos. 
Even the deposits at Kiev, which were substantial enough to encourage 
commercial exploitation,68 cannot have figured prominently. Further­
more, all amber myths indicate that the material known in very early 
times was washed up on the shore (succinum haustile), not dug from 
the ground (succinum fossile). On balance, Professor Marinatos' sea 
route via Lipari to Sandy Pylos and thence to Mycenae and other 
inland and island sites is by far the most convincing. Hachmann has 
established a firm connection between Mycenaean amber ornaments 
and those of the Wessex-Culture,69 but that need not mean that the 
amber used in their manufacture was found in England. Indeed, the 
quantities of amber naturally occurring in England 70 are almost 
certainly insufficient for the large drilled plates Hachmann discusses. 
England thus would seem to have been another way station, and the 

"L. Bernabb Brea and M. Cavalier, II Castello di Lipari e il Museo Archeologico Eoliano 
(Palermo n.d.) 2l. 

65 Beck, supra n.58. 
66 F. T. Koppen, "Vorkommen des Bernsteins in Russland," Petermanns geographische 

Mitteilungen 39 (1893) 249-253. 
67 Janet R. Bacon, The Voyage of the Argonauts (London 1925). 
58 V. Chvojka, quoted by Evans, supra n.37 (a), 44 n.3. 
69 R. Hachmann, "Bronzezeitliche Bernsteinschieber," BayAkadWiss Heft 22 (1957) 1-36. 
70 H. B. Woodward, Geology of England and Wales (London 1887) 450-451 and earlier 

literature cited there. 



210 ANALYSIS AND PROVENIENCE OF AMBER, I 

o 

o 
o 

o o 

o 

o 

Moskva 
o • 

o 

Figure 4. AMBER IN RUSSIA, AFTER KOPPEN 

o 

origin of the raw amber, though not of the workmanship, must be 
sought at least in Jutland and perhaps even in the eastern Baltic. 

The existence of Baltic amber in southern Russia is nevertheless 
significant in that it forbids us to equate the establishment of the sea 
route with the earliest occurrence of Baltic amber in the eastern 



CURT W. BECK 211 

Mediterranean. The oldest and characteristically small finds may well 
have come from the immediate north of the Euxine where. even 
without directed search. a chance find of Baltic amber in a river bed 
must always have been possible. Only with the beginning of veritable 
hoards of large beads need we look for direct importation from the 
north of Europe. 

There is hope that the systematic and exhaustive study of Minoan, 
Mycenaean. and later Greek amber artifacts by infrared spectro­
photometry will help towards the solution of such problems.71 

VASSAR COLLEGE 

May, 1966 

71 We are happy to acknowledge the support of this work by the National Science 
Foundation through Grant GS-739 (Anthropology) for the analysis of amber artifacts and 
through Grant GP-4729 (Earth Sciences) for the classification of naturally occurring fossil 
resins.