Reconstructing Early Cretan Metallurgy:: Analytical Evidence from Kephala Petras, Siteia

Full text

Published by
INSTAP Academic Press
Philadelphia, Pennsylvania
2011
Metallurgy:
Understanding
How,
Learning
whY
Studies in Honor of James D. Muhly
PREHISTORY MONOGRAPHS 29
edited by
Philip P. Betancourt and Susan C. Ferrence
MU_Front_new_Layout 1 5/11/2011 10:13 AM Page iii

Design and Production
INSTAP Academic Press
Printing
CRWGraphics, Pennsauken, New Jersey
Binding
Hoster Bindery, Inc., Ivyland, Pennsylvania
Copyright © 2011
INSTAP Academic Press
Philadelphia, Pennsylvania
All rights reserved
Printed in the United States of America
Library of Congress Cataloging-in-Publication Data
Metallurgy, understanding how, learning why : studies in honor of James D. Muhly / edited by Philip P. Betancourt and
Susan C. Ferrence.
p. cm. -- (Prehistory monographs ; v. 29)
Includes bibliographical references.
ISBN 978-1-931534-57-4 (hardcover : alk. paper)
1. Metal-work, Prehistoric. 2. Bronze age--Cyprus. 3. Bronze age--Aegean Sea. 4. Bronze implements--Cyprus. 5.
Bronze implements--Aegean Sea. 6. Cyrpus--Antiquities. 7. Aegean Sea--Antiquities. 8. Muhly, James David. I.
Betancourt, Philip P., 1936- II. Ferrence, Susan C., 1974-
GN799.M4M48 2011
939'.37--dc23
2011017917
FPO
FSC
MU_Front_new_Layout 1 5/11/2011 10:13 AM Page iv

AKR excavation number, Akrotiri, Thera
cm centimeter
dia. diameter
EBA Early Bronze Age
EC Early Cycladic
EChal Early Chalcolithic
ED-XRF emission dispersive X-ray fluorescence
EH Early Helladic
EM Early Minoan
gr gram
h. height
HM Herakleion Archaeological Museum
HNM Hagios Nikolaos Archaeological
Museum
L. length
LBA Late Bronze Age
LC Late Cycladic or Late Cypriot
LChal Late Chalcolithic
LH Late Helladic
LM Late Minoan
m meter
MBA Middle Bronze Age
MC Midlle Cycladic
MChal Middle Chalcolithic
MH Middle Helladic
MM Middle Minoan
NCSR National Center for Scientific
Research “Demokritos”
NM National Archaeological Museum
of Greece
NMD Neolithic Museum, Diros, Mani
pers. comm. personal communication
pers. obs. personal observation
pres. preserved
Abbreviations for periodicals in the bibliographies of the individual articles follow the conventions of
the American Journal of Archaeology 111.1 (2007), pp. 14–34.
List of Abbreviations
MU_Front_new_Layout 1 5/11/2011 10:13 AM Page xxxi

METALLURGY: UNDERSTANDING HOW, LEARNING WHY
SEM/EDX scanning electron microscopy and
energy dispersive microanalyses
SM Siteia Archaeological Museum
th. thickness
w. width
wt. weight
XRD X-ray diffractometry
th. thickness
XRF X-ray fluorescence spectrometry
xxxii
MU_Front_new_Layout 1 5/11/2011 10:13 AM Page xxxii

Introduction
on Crete. Evidence of similar date is ex tremely rare
both within Crete (Betancourt 2006, 2007) and in
the rest of the Aegean (Nakou 1995; Muhly 2002,
2006; Papadatos 2007a). For this reason, an analyti-
cal project was undertaken by the Laboratory of
Archaeometry of the National Center for Scientific
Research (NCSR) “Demokritos” for the technologi-
cal study of the metallurgical activities evidenced at
Kephala Petras. The project combined microscopic
analysis, chemical analysis, and phase microanalysis
of ore and slag samples in an effort to elucidate the
major technological aspects of the smelting process.
This paper presents the analytical results from the
study of the metallurgical finds and discusses their
implications for our understanding of the early met-
allurgical technology in this area.
Recent excavations at Kephala Petras in eastern
Crete (Papadatos 2008; Papadatos et al., forthcom-
ing) brought to light new evidence for copper-
smelting activities in Crete, dating at least to the ear-
liest part of the Early Minoan (EM) I period or pos-
sibly to the end of the Final Neolithic (FN) period
(Papadatos 2007a). The finds, typical of early
copper-smelting processes, include two pieces of
copper ore, six slag fragments, and four pieces of
iron ore. To these probably are to be added a small
number of deformed clay fragments clearly subject-
ed to high temperatures, which might represent
refractory material used in the metallurgical process.
The importance of these finds, despite their small
quantity, lies in the fact that they are among the ear-
liest direct evidence for copper-smelting activities
8
Reconstructing Early Cretan Metallurgy:
Analytical Evidence from Kephala Petras, Siteia
Mihalis Catapotis, Yannis Bassiakos, and Yiannis Papadatos
CHAPTER
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 69

MIHALIS CATAPOTIS, YANNIS BASSIAKOS, AND YIANNIS PAPADATOS70
Raw Materials
Direct evidence for the nature of the raw materi-
als used in the smelting process at Kephala Petras is
provided by the two small pieces of copper ore
found at the site. Microscopic examination of the
ore samples indicated that they consist of secondary
minerals of copper and iron oxides in a siliceous
matrix, containing only a small amount of residual
pyrite and chalcopyrite (Fig. 8.1), which is also
reflected in the low levels of sulfur in their bulk
composition (Table 8.1). That the smelted ores were
of the oxidized type is further corroborated by the
low levels of sulfur and the limited presence of
matte inclusions in all slag samples.
The smelting of oxidized ores for the production
of copper was standard throughout the southern
Aegean during the EBA, as shown by several studies
of smelting sites on Seriphos, Keros, Kea, Kythnos,
and Crete (Gale et al. 1985; Bassiakos and Doumas
1998; Papastamataki 1998; Bassiakos and Catapotis
2006; Bassiakos and Philaniotou 2007; Catapotis
and Bassiakos 2007; Georgakopoulou 2007). The
principle of the process was the conversion of cop-
per oxides to metallic copper by creating a reducing
atmosphere in the furnace (Newton and Wilson
1942, 151). Copper is a fairly noble metal, and its
oxides could be relatively easily reduced. The major
challenge was rather the removal of the “earthy”
component of the ore (i.e., the gangue minerals) by
creating a fluid slag which could allow the separa-
tion of the reduced metal. This involved firstly the
manipulation of the composition of the charge and
secondly the attainment of sufficiently high temper-
atures during the process.
Unfortunately, the evidence is too limited to pro-
vide a quantifiable picture of the gangue minerals of
the ores smelted at the site. Bulk analysis of the two
copper ore samples proved one of them to be iron
rich and the other rather siliceous (Table 8.1). This
simply could reflect minor variations among differ-
ent pieces collected from the same ore deposit, given
that such deposits are always heterogeneous bodies.
It is equally possible, however, that copper ores
brought to Kephala Petras for smelting (if indeed
more than one smelting event took place there) did
not always have the same mineralogical composi-
tion. Whether this implies the use of ores from dif-
ferent sources is open to further analysis. It remains,
however, an intriguing possibility, given that lead
isotope analysis of slags from the other copper-
smelting site on Crete, namely Chrysokamino,
revealed that ores from various copper deposits were
probably smelted at the site (Stos-Gale and Gale
2006, 313–316).
Whatever the case, the chemical composition of
the slag from the site reveals that some care was
probably taken to ensure that the composition of the
charge would enable the formation of a liquid slag
during the smelt. Indeed, as seen in Figure 8.2, the
chemical composition of all five slags falls into an
area of the FeOx-SiO2-CaO-Al2O3phase diagram
characterized by relatively low liquidus tempera-
tures in the range of 1250oC. Moreover, it is notable
that the samples form a fairly tight compositional
cluster, which contrasts with the variability reflect-
ed in the two copper ores. Although we should not
dismiss the possibility that all slag samples come
from a single smelt, this consistency might suggest
that smelters at Kephala Petras exerted some control
over the composition of the smelting charge.
This control could have been achieved through a
range of possible beneficiation and fluxing tech-
niques. As far as beneficiation is concerned, the
small size of the copper ore fragments (1.6–2.8 cm)
and the large concentrations of copper oxides clear-
ly visible in their surface (see also Table 8.1) may
suggest that some sort of beneficiation process was
taking place prior to smelting, involving the crushing
of the ore and the selection of the richest pieces for
the smelt. As for the possibility of fluxing, the pieces
of iron ore found at the site may be illuminating. Of
the four iron ores discovered during the excavations,
only one has been analyzed so far. Arguably, its
chemical composition (Table 8.1) does not seem
ideal for the fluxing of either siliceous or ferrous
copper-ores. Nonetheless, the fact that it contains no
traces of copper, coupled with its distinctively high
levels of alumina, suggests that this piece of ore
probably came from a different deposit from that of
the copper ores found at the site. Moreover, that the
iron ore was found in the same context with a piece
of siliceous copper ore suggests that it might have
been somehow related to the metallurgical activities
at the site. Put together, this evidence may raise the
possibility that this piece of iron ore is an example
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 70

RECONSTRUCTING EARLY CRETAN METALLURGY 71
Figure 8.1. Copper-ore sample
KP 03/1156. Note the residual
sulphide min erals seen as
white in the micro-photograph
(optical microscope; cross-
polarized light [XPL]).
Find
number Type Percent (%)
FeO CuO SiO2Al2O3MgO CaO K2OTiO2BaO SO3
KP 03/473 copper ore 48 39 6.5 0.4 0.5 0.4 0.1 — — 5.9
KP 03/1156 copper ore 16 48 34 —0.3 0.4 0.1 — — 1.7
KP 03/855 iron ore 36 —35 27 —0.4 0.3 1.4 — —
KP 03/1119 slag 48 2.2 32 5.4 1.3 7.5 1.2 —3.2 0.3
KP 03/230 slag 50 2.9 35 4.2 1.2 6.4 0.7 — — 0.3
KP 03/244 slag 50 1.3 32 4.6 1.3 9.0 1.1 0.2 — —
KP 03/759 slag 27 3.8 45 9.4 1.8 9.0 1.7 —1.8 —
KP 03/194 slag 43 1.6 37 5.7 1.8 9.0 1.2 0.5 — —
Table 8.1. Chemical composition from area scans of ore and slag samples from Kephala Petras using scanning
electron microscopy–energy dispersive spectroscopy (SEM-EDS) (multiple scans; normalized data).
Figure 8.2. Reduced chemical
composition of slag samples from
Kephala Petras plotted on the
Fe/SiO2–CaO (+7% Al2O3) phase
diagram (diagram after Kongoli
and Yazawa 2001, fig. 11).
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 71

MIHALIS CATAPOTIS, YANNIS BASSIAKOS, AND YIANNIS PAPADATOS72
(though not representative) of minerals that were
added to the charge, when necessary, to adjust its
composition in order to ensure the formation of a
liquid slag. In other words, although further research
is necessary, it seems possible that fluxes were used
in the smelting process at Kephala Petras.
Smelting Conditions
To seek evidence for the temperature and redox
conditions that prevailed during the smelting pro cess
at Kephala Petras, the mineralogy and texture of the
slag were examined. Macroscopic and microscopic
examination of five samples indicated that although
the slag is not fully liquified it contains only a few
pieces of unreacted raw materials, such as copper ore
and, more commonly, silica (Fig. 8.3). This suggests
that the maximum temperatures attained during the
smelt were close to, though not necessarily above,
the liquidus temperature of the slag, which has been
estimated to be about 1250oC.
That the maximum smelting temperature was
close to the liquidus temperature of the slag is also
demonstrated by the notable levels of copper in the
slag. More specifically, the Cu weight-% in all five
samples is above 1%, reaching a maximum of 3% in
the more siliceous slag (Table 8.1). This suggests
that the smelting temperature was not high enough
to enable the production of a fluid slag. The slag
remained instead fairly viscous, leading to the
entrainment of metallic prills and thus increasing
copper losses.
Evidence for the redox conditions during the
smelt is offered by the oxidation state of copper and
iron as well as the composition of metal-rich prills
in the slag. The virtual absence of primary (i.e., not
post-depositional) cuprite and the copper content of
the slag (1%–3%) suggests that the partial pressure
of oxygen (pO2) did not exceed the 10-5atm. On the
other hand, the low levels of iron in the copper prills
(0.4%–3.0%) and matte inclusions (1.1%–4.0%)
(Table 8.2), coupled with the predominance of mag-
netite (against wustite) in all slag samples, argue
against pO2values below the 10-11atm. Overall, the
mineralogy of the slag samples from Kephala Petras
points to moderate redox conditions, with the pO2
ranging from 10-6 to 10-11atm.
What is, perhaps, more important is that the min-
eralogy of the slag samples reflects significant vari-
ability in the redox conditions prevailing during the
smelt. Although the small size of the sample does
not allow firm conclusions, the co-existence of
delafossite, magnetite, and wustite in a single slag
sample (Fig. 8.4) strongly suggests that the smelting
conditions were far from stable. Similar conclu-
sions can be drawn from the highly variable levels
of iron in copper prills (Table 8.2). Considering also
the small number and size of the slag pieces found
at the site and the absence of any evidence for a
major metallurgical installation, this variability
could be taken to imply that the smelting process
was taking place inside a small ceramic container,
possibly a bowl-shaped crucible, or inside a small
hearth, lined with refractory clay.
We could consider, at this point, the deformed
clay fragments found at the site, whose state of vitri-
fication suggests that they have been subjected to
high temperatures (Papadatos 2007a, figs. 10.7,
10.8). Although none of the sherds bears traces of
copper or slag, given the absence of any fire-destruc-
tion layers at the site, it is very likely that these
burned ceramics may relate to a pyrotechnological
process. If this is indeed the case, it is possible that
they are the remains of small ceramic vessels or
hearths used as crucibles for copper smelting. This is
reinforced further by petrographic analysis, which
shows that they are made of a semicoarse fabric con-
taining much organic temper and rare inclusions of
quartz and quartzite (E. Nodarou, pers. comm.)—a
fabric not used for any other classes of pottery at the
site. Moreover, fabrics of this type with organic tem-
per were very common for the manufacture of
refractory material used for metallurgical purposes
in Early Minoan Crete (Doonan, Day, and
Dimopoulou-Rethemiotaki 2007, 104–105).
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 72

RECONSTRUCTING EARLY CRETAN METALLURGY 73
Find number Phase
Percent (%)
Cu Fe S
KP 03/1119 matte inclusion 77 4.0 19
KP 03/1119 matte inclusion 81 1.1 18
KP 03/1119 copper prill 99 0.5 0.3
KP 03/1119 copper prill 98 1.9 0.2
KP 03/230 copper prill 97 3.0 0.4
KP 03/244 copper prill 98 1.6 0.2
KP 03/759 copper prill 99 0.5 0.6
KP 03/194 copper prill 99 0.4 0.2
Table 8.2. Chemical composition of metallic inclusions in slag samples from Kephala Petras determined by SEM-EDS
(normalized data).
Figure 8.3. Slag sample KP 03/230
containing piece of unreacted copper-ore
(dark inclusion at the center) surrounded
by magnetite skeletons (optical micro -
scope with plain-polarized light [PPL]).
Figure 8.4. Slag sample KP 03/1119 that
features the co-presence of delafossite
laths (center), magnetite skeletons (left),
and wustite dendrites (right) (optical
micro scope with PPL).
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 73

MIHALIS CATAPOTIS, YANNIS BASSIAKOS, AND YIANNIS PAPADATOS74
Product
Analysis of the metallic phases embedded in the
slag samples suggests that the product of the smelt-
ing activities evidenced at Kephala Petras was
metallic copper containing low levels of iron and
possibly some copper-rich matte inclusions. Arsenic
and other impurities were below the detection limit
of the analytical instrument, as was the case with the
two copper ore samples from the site. Such a prod-
uct is compatible with the technology reconstructed
at the site, i.e., with the use of oxidized copper ores
containing only small amounts of residual sulfides
and the attainment of moderate redox conditions
during the smelt. A final refining stage would have
been necessary in order to remove iron and other
impurities before using the metal for the manufac-
ture of artifacts.
This new evidence strengthens the view that the
predominance of arsenical copper in the metalwork
of the Early Bronze Age (EBA) southern Aegean
should not be attributed to a widespread practice
of smelting arsenic-rich raw materials. Indeed, as
previous studies have shown, most copper-smelting
sites in that area produced arsenic-free copper
(Bassiakos and Doumas 1998; Papastamataki
1998; Bassiakos and Philaniotou 2007). Arsenical-
copper smelting is only evidenced at the sites of
Chrysokamino on Crete (Catapotis and Bassiakos
2007); Daskaleio-Kavos on Keros, where arsenic-
free copper and arsenical copper were produced
by two distinct processes (Georgakopoulou 2007);
and Skour ies on Kythnos, where a small number of
slag samples were found to contain arsenic-rich
copper prills (Gale et al. 1985; contra Bassiakos
and Philaniotou 2007, 51–52).
Kephala Petras in the Context of Early Cretan Metallurgy
Kephala Petras and Chrysokamino
Despite the limited available evidence, it would be
worth putting the results of our analysis within the
wider context of early Cretan metallurgy. Interesting
conclusions can be drawn from a technological com-
parison of the evidence from Kephala Petras with its
sole counterpart on Crete, the copper-smelting site
of Chrysokamino. As demonstrated in Table 8.3, the
copper-smelting technologies employed at Kephala
Petras and Chrysokamino bear similarities but also
significant differences.
At both sites, oxidized copper ores were used, and
the composition of the charge seems to have been
controlled, to a lesser or greater extent, through
some combination of ore selection, beneficiation,
and/or fluxing. What differentiated Chrysokamino,
however, was the use of arsenic-rich raw materials,
which enabled the repeated production of arsenical
copper (Catapotis and Bassiakos 2007).
The smelting process at Chrysokamino was con-
ducted in furnaces consisting of a perforated conical
shaft (a technology also evidenced in the western
Cyclades and Attica) laid on top of a clay-lined
pit, which probably enabled the smelting of tens of
kilograms of ore in every single smelt. At Kephala
Petras there is no direct evidence for the type of ves-
sel used in the smelting process. However, the size,
texture and mineralogy of the analyzed slag samples
seem to suggest the use of a small clay-lined hearth
or a crucible, which would have enabled the smelt-
ing of some hundreds of grams of ore in each smelt-
ing episode. The difference in the capacity of the
smelting installations and the quantity of the metal-
lurgical debris from the two sites may reflect the
difference between regular processing of copper
ores at an activity-specific location (Chrysokamino)
on the one hand, and occasional small-scale smelts
conducted very close to or within the borders of a
settlement (Kephala Petras) on the other.
At Chrysokamino, oxygen was supplied into the
system in two ways. First, wind penetrated through
the numerous perforations on the shaft, enabling
the early combustion of the fuel and the efficient
preheating of the charge (Catapotis, Pryce, and
Bassiakos 2008). At the lower levels of the furnace
air was supplied by pot-bellows, evidence of which
has been reported from the site (Betancourt 2006,
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 74

RECONSTRUCTING EARLY CRETAN METALLURGY 75
Establishing the chronology of the metallurgical
activities at Chrysokamino presents greater diffi-
culties. The metallurgical evidence derives from an
extensive slag heap, which constituted a disturbed,
unstratified deposit, containing a mixture of FN,
EM I–II, and mostly EM III–MM IA pottery
(Betancourt 2006, 68–71). Only a small area of this
extensive deposit was found sealed by the succes-
sive floors of an EM III–MM IA hut. Therefore, the
EM III–MM IA floor is the secure terminus ante
quem for the dating of the slag heap. Based on this
evidence, it is possible that metallurgical activity
started as early as the FN (Betancourt 2006, 179),
but the main period of use seems to be EM III
(Muhly 2006, 155). Further support comes also
from the pot bellows, which can be securely dated
solely to the EM III–MM IA period (Betancourt
2006, 126).
Whatever the case, the fact that two copper-
smelting sites located on the northern shores of
eastern Crete demonstrate such different smelting
techniques is indicative of the technological vari-
ability that characterized early metallurgical prac-
tices in the southern Aegean during the 3rd
millennium B.C. Similar variability has been also
noticed in the Cyclades, sometimes within the same
island, e.g., Kythnos (Bassiakos and Philaniotou
2007) and Seriphos (Georgakopoulou 2005) or even
the same settlement, e.g., Kavos North on Keros
(Georg a ko poulou 2004, 2007).
Smelting Processes Chrysokamino Kephala Petras
Copper ores Oxidized Oxidized
Beneficiation Ye s? Yes?
Fluxing Iron ore? Iron-ore?
Arsenical ingredient Yes —
Smelting vessel Perforated shaft furnace Crucible?
Air supply Wind + pot-bellows ?
Furnace temperature ≈1300oC<1250oC
Redox conditions Moderate Variable
Slag processing Crushing, pulverizing(?) ?
Copper losses 1% 1%–3%
125–132). The temperature and redox atmosphere
thus created inside the furnace were sufficient to
ensure the reduction of most copper-oxide to its
metallic state, the formation of a fluid slag, and a
good separation of the two formed phases. In the
case of Kephala Petras, the manner of air-supply in
the process remains uncertain, but smelting condi-
tions appear to have been variable, and the temper-
atures attained were not sufficient to lower the
viscosity of the slag, resulting in a less successful
slag-metal separation compared to Chrysokamino.
Nonetheless, it should be noted that copper losses
in the slag from Kephala Petras (1%–3% in ana-
lyzed samples) are still of the same order of mag-
nitude as those of Chry so kamino and other
smelting sites in the Aegean (Catapotis and
Bassiakos 2007, fig. 4.7).
Assigning a chronological significance to the
differences evidenced at the two sites, though
tempting, is not without problems. Direct metallur-
gical evidence at Kephala Petras (ores and slag)
comes from secondary deposits, containing FN and
EM IA pottery, which were sealed by the floors of
an EM IA building. Only the burned clay fragments
come from undisturbed pure FN deposits, but their
association with metallurgy, though possible, has
not yet been firmly established. Thus, the current
evidence cannot provide a precise dating for metal-
lurgy at Kephala Petras; it can only set a terminus
ante quem to the EM IA period (Papadatos 2007a).
Table 8.3. Comparison of technical aspects of the smelting processes at Kephala Petras and Chrysokamino (Catapotis
and Bassiakos 2007).
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 75

MIHALIS CATAPOTIS, YANNIS BASSIAKOS, AND YIANNIS PAPADATOS76
Kephala Petras and the Kampos Group
Metallurgical Sites
One of the most important features of the early
part of the EBA in Crete is the appearance on the
northern shore of sites with strong Cycladic affilia-
tions, namely Poros-Katsambas, Hagia Photia, and
Gournes (Davaras and Betancourt 2004; Galanaki
2006; Wilson, Day, and Dimopoulou-Rethemiotaki
2008). Known also as the Kampos Group in the
EBA Cyclades, they date to the EM IB–IIA period.
Recent approaches associate these sites with the
emergence of nodal points controlling the inflow
of exotic materials and techniques of Cycladic
origin, including metals and metallurgy, to Crete
(Day, Wilson, and Kiriatzi 1998; Broodbank 2000,
300–304; Papadatos 2007b; Wilson, Day, and
Dimopoulou-Rethemiotaki 2008).
Two sites, Poros-Katsambas and Hagia Photia,
have yielded metallurgical evidence. At Poros-
Katsambas, a rich metalworking assemblage has
been discovered, including crucibles, molds, tuyères,
copper ingots, arsenic-bearing minerals, casting
spills, melting slag, and dross (Doonan, Day, and
Dimopoulou-Rethemiotaki 2007). At Hagia Photia
the material comes from two different contexts. The
cemetery has yielded copper-based artifacts and two
complete crucibles, which, according to preliminary
analyses, were used for the melting of copper
(Betancourt and Muhly 2007, 150). At the adjacent
settlement, nine chisel molds have been discovered
in an EM IB–IIA deposit underlying the Middle
Minoan strata (Tsipopoulou 2007).
What is important to note is that at both sites the
evidence relates to metalworking, not to the produc-
tion of metal from its ores. In other words, the evi-
dence suggests that metal from the Cyclades and/or
other sources, reached the “Kampos Group” sites in
Crete in its ready-made form. The evidence from
Kephala Petras shows, instead, that copper ore was
brought to the settlement in order to be smelted for
the extraction of the metal. Whether this should be
taken to reflect a different mode of metal procure-
ment (Papadatos 2007a, 163–165) in action on Crete
before the Kampos Group phase remains to be
addressed by future research. In any case, this con-
trast reinforces the picture of variability in the spatial
arrangement of metallurgical activities evidenced in
the archaeological record of the Early Bronze Age in
the southern Aegean (Catapotis 2007).
Conclusions
Despite the limited amount of metallurgical evi-
dence from Kephala Petras, it was possible to
reconstruct a copper-smelting technology that is
compatible with what is known so far about early
smelting techniques in the southern Aegean. The
process involved the smelting of oxidized ores,
possibly mixed with a flux, inside a small clay
container where fairly high temperatures and
moderate redox conditions were attained. This
enabled the production of copper with low levels
of iron that were probably removed in a final
refining stage, not attested so far at Kephala.
Although the evidence from the site seems to
point to small-scale smelting activities, it proves
beyond doubt that copper production was taking
place on Crete before the “Kampos Group” period.
Therefore, the paucity of metal objects and their
absence from tombs of the FN–EM IA period,
such as Pseira (Betancourt and Davaras, eds.,
2003), Partira (Mortzos 1972), and Hagios
Nikolaos–Palaikastro (Tod 1902–1903), should
not be viewed as evidence for the non-practicing of
metallurgy on Crete during that time. Rather, it is
indicative of different depositional practices,
which have been convincingly associated with the
changing role of metal in the southern Aegean
societies during the EB I–II period (Nakou 1995;
Papadatos 2007a).
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 76

RECONSTRUCTING EARLY CRETAN METALLURGY 77
Acknowledgments
Finally, we thank the anonymous reviewer for use-
ful comments on an earlier version of this paper.
The study related to this paper took place as part of
the project “Study and Publication of the FN–EM I
Settlement at Kephala Petras, Siteia,” funded by the
Institute for Aegean Prehistory (INSTAP).
The authors would like to thank the excavator of
Petras, Dr. Metaxia Tsipopoulou, for granting us
permission to study and publish the material from
Kephala. We would also like to thank the director of
the 24th Ephorate of Prehistoric and Classical
Antiquities, Vili Apostolakou, for permissions to
sample and study the metallurgical evidence.
References
Bassiakos, Y., and M. Catapotis. 2006. “Reconstruction
of the Copper Smelting Process at Chrysokamino
Based on the Analysis of Ore and Slag Samples,” in
Betancourt, 2006, pp. 329–353.
Bassiakos, Y., and C. Doumas. 1998. “The Island of
Keros and Its Enigmatic Role in the Aegean EBA: A
Geoarchaeological Approach,” in Αργυρίτις Γη, A.
Fragiskos, ed., Athens, pp. 55–64.
Bassiakos, Y., and O. Philaniotou. 2007. “Early Copper
Production on Kythnos: Archaeological Evidence
and Analytical Approaches to the Reconstruction of
Metallurgical Process,” in Day and Doonan, eds.,
2007, pp. 19–56.
Betancourt, P.P. 2006. The Chrysokamino Metallurgy
Workshop and Its Territory (Hesperia Suppl. 36),
Princeton.
———. 2007. “The Final Neolithic to Early Minoan III
Metallurgy Site at Chrysokamino, Crete,” in Day and
Doonan, eds., 2007, pp. 57–67.
Betancourt, P.P., and J.D. Muhly. 2007. “The Crucibles
from the Aghia Photia Cemetery,” in Day and
Doonan, eds., 2007, pp. 146–153.
Betancourt, P.P., and C. Davaras, eds. 2003. Pseira VII:
The Pseira Cemetery 2. Excavation of the Tombs
(Prehistory Monographs 6), Philadelphia.
Broodbank, C. 2000. An Island Archaeology of the Early
Cyclades, Cambridge.
Catapotis, M. 2007. “On the Spatial Organisation of
Copper Smelting Activities in the Southern Aegean
during the Early Bronze Age,” in Day and Doonan,
eds., 2007, pp. 207–223.
Catapotis, M., and Y. Bassiakos. 2007. “Copper
Smelting at the Early Minoan Site of Chrysokamino
on Crete,” in Day and Doonan, eds., 2007, pp. 68–83.
Catapotis, M., O. Pryce, and Y. Bassiakos. 2008.
“Preliminary Results from an Experimental Study of
Perforated Copper-Smelting Shaft Furnaces from
Chrysokamino (Eastern Crete),” in Aegean Metallurgy
in the Bronze Age: Proceedings of an International
Symposium Held at the University of Crete, Reth ym -
non, Greece on November 19–21, 2004, I. Tzachili, ed.,
Athens, pp. 113–122.
Davaras, C., and P.P. Betancourt. 2004. The Hagia Photia
Cemetery I: The Tomb Groups and Architecture
(Prehistory Monographs 14), Philadelphia.
Day, P.M., and R.C.P. Doonan, eds. 2007. Metallurgy in
the Early Bronze Age Aegean (Sheffield Studies in
Aegean Archaeology 7), Oxford.
Day, P.M., D.E. Wilson, and E. Kiriatzi. 1998. “Pots,
Labels and People: Burying Ethnicity in the Cemetery
of Aghia Photia, Siteias,” in Cemetery and Society in
the Bronze Age (Sheffield Studies in Aegean Archaeo -
logy 1), K. Branigan, ed., Sheffield, pp. 133–149.
Doonan, R.C.P., P.M. Day, and N. Dimopoulou-
Rethemiotaki. 2007. “Lame Excuses for Emerging
Complexity in Early Bronze Age Crete,” in Day and
Doonan, eds., 2007, pp. 98–122.
Galanaki, K. 2006. “Πρωτομινωικό ταφικό σύνολο στην
πρώηv Αμερικανική Βάση Γουρνών Πεδιάδοϛ,” in
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 77

MIHALIS CATAPOTIS, YANNIS BASSIAKOS, AND YIANNIS PAPADATOS78
Πεπραγμένα Θʹ Διεθνούϛ Κρητολογικού Συνε -
δρίου Α(2), Herakleion, pp. 227–242.
Gale, N.H., A. Papastamataki, Z.A. Stos-Gale, and K.
Leonis. 1985. “Copper Sources and Copper Metal -
lurgy in the Aegean Bronze Age,” in Furnaces and
Smelting Technology in Antiquity (British Museum Oc -
casional Paper 48), P. Craddock and M. Hughes, eds.,
London, pp. 81–93.
Georgakopoulou, M. 2004. “Metallurgy at the Early
Bronze Age Site of Daskaleio-Kavos on the island of
Keros (Cyclades, Greece),” Institute of Archa eo metal -
lurgical Studies Newsletter 24, pp. 3–12.
———. 2005. Technology and Organization of Early
Cycladic Metallurgy: Copper on Seriphos and Keros,
Greece, Ph.D. diss., University College London.
———. 2007. “Metallurgical Activities within Early
Cycladic Settlements: The Case of Daskaleio-Kavos,”
in Day and Doonan, eds., 2007, pp. 123–134.
Kongoli, F., and A. Yazawa. 2001. “Liquidus Surface of
FeO-Fe2O3-SiO2-CaO Slag Containing Al2O3, MgO,
and Cu2O at Intermediate Oxygen Partial Pressures,”
Metallurgical and Materials Transactions 32B, pp.
583–592.
Mortzos, C. 1972. “Πάρτιρα, μία πρώϊμος μινωϊκὴ
ὁμάϛ,” Ἐπετηρίς Ἐπιστημονικῶν Ἐρευῶν 3, pp.
386–421.
Muhly, J.D. 2002. “Early Metallurgy in Greece and
Cyprus,” in Anatolian Metal II (Der Anschnitt 15),
Ü. Yalçin, ed., Bochum, pp. 77–82.
———. 2006. “Chrysokamino in the History of Early
Metallurgy,” in Betancourt 2006, pp. 155–177.
Nakou, G. 1995. “The Cutting Edge: A New Look at
Early Aegean Metallurgy,” JMA 8 (2), pp. 1–32.
Newton, J., and C.L. Wilson. 1942. Metallurgy of Cop -
per, New York.
Papadatos, Y. 2007a. “The Beginning of Metallurgy in
Crete: New Evidence from the FN–EM I Settlement
at Kephala Petras, Siteia,” in Day and Doonan, eds.,
2007, pp. 154–167.
———. 2007b. “Beyond Cultures and Ethnicity: A
New Look at Material Culture Distribution and
Inter-regional Interaction in the Early Bronze Age
Southern Aegean,” in Mediterranean Crossroads, S.
Antoniadou and A. Pace, eds., Athens, pp. 419–451.
———. 2008. “The Neolithic–Early Bronze Age
Transition in Crete: New Evidence from the
Settlement at Petras Kephala, Siteia,” in Rethinking
the Cretan Neolithic (Sheffield Studies in Aegean
Archaeology 8), V. Isaakidou and P.D. Tomkins, eds.,
Oxford, pp. 261–275.
Papadatos, Y., P. Tomkins, E. Nodarou, and I. Iliopoulos.
Forthcoming. “The Beginning of Early Bronze Age in
Crete: Continuities and Discontinuities in the Ceramic
Assemblage at Kephala Petras, Siteia,” in The Aegean
Early Bronze Age: New Evidence, C. Doumas, A.
Gannikouri, and O. Kouka, eds., Athens.
Papastamataki, A. 1998. “Μεταλλουργικές δραστηρι -
ότητες στην Κέα κατά την Αρχαιότητα, νεότερα
στοιχεία,” in Κέα-Κύθνος: Ιστορία και Αρχαιο -
λογία (MELETEMATA 27), L.G. Mendoni and A.
Mazarakis-Ainian, eds., Athens, pp. 759–764.
Stos-Gale, Z.A., and N.H. Gale. 2006. “Lead Isotope
Analyses of Slags from Chrysokamino,” in
Betancourt 2006, pp. 299–319.
Tod, M.N. 1902–1903. “Excavations at Palaikastro, II.
Hagios Nikolaos,” BSA 9, pp. 336–343.
Tsipopoulou, M. 2007. “Aghia Photia–Kouphota: A
Centre for Metallurgy in the Early Minoan Period,”
in Day and Doonan, eds., 2007, pp. 135–145.
Wilson, D.E., P.M. Day, and N. Dimopoulou-
Rethemiotaki. 2008. “The Gateway Port of Poros-
Katsambas: Trade and Exchange between
North-Central Crete and the Cyclades,” in Horizon: A
Colloquium on the Prehistory of the Cyclades, N.
Brodie, J. Doole, G. Gavalas, and C. Renfrew, eds.,
Cambridge, pp. 261–270.
MU_CH8_Catapotis_Bassiakos_Papadatos_new_Layout 1 5/5/2011 1:31 PM Page 78