Technology issues of Byzantine glazed pottery from Corinth, Greece

E. Palamara, N. Zacharias, M. Xanthopoulou, Zs. Kasztovszky, I. Kovács, D.

Palles, E.I. Kamitsos

PII: S0026-265X(16)30078-9
DOI: doi: 10.1016/j.microc.2016.06.008
Reference: MICROC 2498

To appear in: Microchemical Journal

Received date: 12 December 2015

Revised date: 31 May 2016
Accepted date: 3 June 2016

Please cite this article as: E. Palamara, N. Zacharias, M. Xanthopoulou, Zs. Kasztovszky,
I. Kovács, D. Palles, E.I. Kamitsos, Technology issues of Byzantine glazed pottery from
Corinth, Greece, Microchemical Journal (2016), doi: 10.1016/j.microc.2016.06.008

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Technology issues of Byzantine glazed pottery from Corinth, Greece

E. Palamara1, N. Zacharias1, M. Xanthopoulou1, Zs. Kasztovszky2, I. Kovács3, D. Palles4

and E.I. Kamitsos4

PT
1
Department of History, Archaeology and Cultural Resources Management, University of the Peloponnese,
24100 Kalamata, Greece
2
Nuclear Analysis and Radiography Department, MTA, Centre for Energy Research, Konkoly Thege

RI
Miklós út 29-33, H-1121 Budapest, Hungary
3
Wigner Research Centre for Physics, Institute for Particle and Nuclear Physics, Konkoly-Thege Miklós út

SC
29-33, H-1121 Budapest, Hungary
4
Theoretical and Physical Chemistry Institute, National Hellenic Research Foundation, 11635 Athens,
Greece

NU
Abstract
A large assemblage of Byzantine glazed pottery from ancient Corinth, Greece was
MA
analyzed by a multi-technique analytical approach in order to determine the production
technology. The samples cover a long time period (10th – 14th c. AD) and a wide range of
the most representative wares and classes. SEM/EDS, a standard non-invasive
microscopy and analytical technique, was applied in combination with Raman
D

spectroscopy and assisted with advanced, high precision techniques (PGAA, milli-PIXE)
TE

in order to examine the ceramic body, the glaze and the clay-glaze interface of the
samples. Parameters of the manufacturing process, such as the selection of clay sources,
the glaze recipe, the glaze application technique and the glaze firing temperature, were
P

examined using a combination of statistical tools and methodologies. This work

comprises the first large scale physicochemical analysis of Byzantine glazed ceramics
CE

from Greece, providing information on the composition and technology of all of the
major typological categories of Medieval Greek glazed ceramics. Furthermore, it
highlights the significant socioeconomic changes that occurred at the -beginning of the
AC

13th c. AD in the Byzantine Empire and their consequences in the manufacturing and
distribution system of glazed ceramics.

Keywords: Byzantine glazed ceramics, Corinth, manufacturing process

1. Introduction
Byzantine glazed pottery has been extensively studied from the point of view of
decoration, shape and macroscopic description of fabric. Morgan [1] was the first to offer
a chronology of the different classes of glazed pottery, which was later revised by
Sanders [2, 3, 4], focusing mainly on the 12th to the mid-13th century. It is generally
accepted that several 12th-13th c. classes, such as ‘Green and Brown Painted’, ‘Slip
Painted’, ‘Fine Sgraffito’, ‘Incised Sgraffito’ and ‘Champlevé’, were produced by a few
specialized workshops and then distributed in the entire Mediterranean region. Indeed,
large quantities of table wares belonging to these classes have been found in numerous
sites in Italy, France, Serbia, Greece, Cyprus, Turkey, Israel, Lebanon etc., but also in
various shipwrecks in the Mediterranean and the Black Sea [5, 6].
 ACCEPTED MANUSCRIPT

Recent studies on Byzantine glazed pottery have mainly focused on the identification of
the major workshops and the shift of the production centers in the 13th c. AD [5, 7, 8, 9,
10]. However, little attention has been given so far to the manufacturing processes
applied and their importance for the overall understanding of the ceramic production of
the period. Such an effort was made by Armstrong et al. [11], however the sample set of

PT
glazed pottery samples examined was very small, thus making it difficult to draw more
general conclusions. More recently, Davis and Stocker [12] examined the typology of a
small assemblage of medieval glazed pottery from Englianos, in Southern Peloponnese,

RI
discussing also the results of the chemical composition of the glazes.
The present work focuses on the application of different analytical techniques and
statistical tools in order to understand and evaluate the manufacturing process for the

SC
production of Byzantine glazed pottery. Methodologies that have been previously applied
to other types of glazed ceramics were now applied for the first time to a Byzantine
assemblage. The combined study of individual parameters of the manufacturing process

NU
led to a thorough understanding of the technology used for the production of Byzantine
glazed ceramics; the resulting ‘chaîne opératoire’ information provides valuable insight
into the social structure of the major Byzantine cities of Greece, through the turbulent
MA
period of the 12th and 13th c. AD.

2. Materials and methods

The assemblage consists of 71 Byzantine glazed pottery fragments, which cover a wide
D

range of the most representative wares and classes; 4 unglazed fragments of cooking
TE

pottery were also analyzed for comparison. All fragments come from unstratified
contexts excavated at Corinth, Greece by the American School of Classical Studies in
Athens and are now part of the study collection of the ceramic teaching collection of the
P

Department of History, Archaeology, and Cultural Resources Management of the

University of the Peloponnese (Kalamata). Most fragments are red wares, except for 17
CE

white ware fragments. Based on their typological characteristics (they all belong to well-
defined and well-dated types, see supra), they can be dated from the 10th to the 14th c.
AD, although the majority dates to the 12th and 13th c. AD (Table 1).
AC

Table 1. Date, ware and style type of the Byzantine glazed pottery sherds

A multi-technique approach was followed for the thorough examination of the clay, the
glaze and the clay-glaze interface (Table 2). The combined use of different techniques
enables a more comprehensive study of the archaeological material, especially when
examining complex, multi-layered items such as glazed ceramics [10, 13, 14].
The ceramic body of all samples was first examined using Fiber Optics Microscopy
(FOM) aiming at a preliminary distinction into groups according to the macroscopic and
microscopic characteristics of the clay. The chemical composition of the clay was
determined by a Scanning Electron Microscope (SEM) by JEOL (JSM-6510LV) coupled
with EDS (Oxford Systems). The analytical data were obtained by INCA software. The
analyses were conducted at 20 kV accelerating voltage, under a magnification of x100
and with a count time of 300 sec. At least 3 measurements were taken for each sample on
different areas of polished fresh cuts. The SEM/EDS composition of all ceramic bodies is
presented in Table A1 of the Appendix.
 ACCEPTED MANUSCRIPT

Representative clay samples were further analysed by Prompt Gamma Activation

Analysis (PGAA) for their bulk elemental composition, including all major and some
trace elements (Table A2 of the Appendix). The PGAA measurements have been
performed on a horizontal cold neutron beam of 9.6·107 neutron/cm2·s-1, at the Budapest
Neutron Centre. The present status of the Budapest PGAA facility is described by

PT
Szentmiklósi et al. [15]. The method is based on the radiative neutron capture, the
composition is calculated on the basis of our PGAA library, using the k0-method [16]. It
is important to mention that PGAA is a non-invasive technique, which has been

RI
previously successfully applied to study the composition of archaeological ceramics [17].
The induced radioactivity decays fast after the investigation and the technique generally
does not require sampling. However, in this particular case it was decided to remove the

SC
glaze from the clay fragments under study, in order to avoid errors related to the presence
of successive layers of glaze and clay.
Since the cold neutron can go through more cm of silicate material, i.e. the ceramics

NU
body, one can obtain the average composition for the irradiated volume. The major
components determined by PGAA are the following: Na2O, MgO, Al2O3, SiO2, P2O5,
SO3, K2O, CaO, TiO2, MnO, Fe2O3, H2O and occasionally CO2. The concentrations of
MA
major oxides are calculated from the concentrations of the elements, based on their
oxidation states. Furthermore, the following trace elements were possible to quantify by
PGAA: B, Cl, Sc, V, Cr, Ba, Nd, Sm and Gd.
SEM/EDS and PGAA show generally good agreement, with a variance of approximately
D

5% between the values of the two techniques for all major elements. In only a few cases
TE

the variance reached 10%. The overestimation of CaO by SEM/EDS in a couple of

samples could be related to the relatively coarse and heterogeneous fabric of these
samples.
P

Additionally, a preliminary mineralogical examination of a few representative clay

samples was carried out by Raman spectroscopy. Raman spectra were obtained by a
CE

Renishaw high confocal micro-Raman spectrometer, using the 488 and 514 nm lines of
an argon ion laser and the 633 nm line of a He-laser.
AC

Table 2. Characteristics of the techniques applied for the examination of clay, glaze and clay-glaze
interface

The bulk chemical composition of glazes was determined by SEM/EDS, which has been
widely applied for the non-destructive examination of archaeological glasses and glazes
[18]. A total of 68 glazes were examined using the same settings described for the clay
(Table A3 of the Appendix). Polished cross sections of the glazes were prepared, in order
to avoid corrosion alterations of the outer layer, as well as picking up signal from the clay
body.
Further analysis by Proton Induced X Ray Emission (PIXE) on representative glaze
samples lead to the qualitative identification of the colourants present in trace levels
(Table A4 of the Appendix). External milli-beam PIXE is a widely used method for non-
destructive elemental analysis of archaeological and cultural heritage objects [19, 20].
The PIXE measurements were carried out at the external millibeam PIXE facility of the
Wigner Research Centre for Physics of the Hungarian Budapest, Hungary. The
collimated proton beam of 2.5 MeV energy was extracted to air through a 7.5 micrometer
thick Kapton foil. The object to be analysed was positioned at a distance of 10 mm from
 ACCEPTED MANUSCRIPT

the exit window by a 3D positioning platform and then carefully oriented. The incident
proton beam was perpendicular to the sample surface. The diameter of the extracted
proton beam on the sample was about 1 mm, the estimated beam current was 2-4 nA. The
energy loss of proton beam passing the Kapton foil and 10 mm of air was calculated by
SRIM code [21], the final energy on the sample was 2.2 MeV. The characteristic X-rays

PT
produced were collected by an Amptek X-123 spectrometer. The SDD type detector of 25
mm2 x 500 µm active volume positioned at 135° with respect to the beam direction at a
distance of 26 mm. The energy resolution of the spectrometer was 130 eV for the Mn-Ka
line. A policarbon absorber of 60 μm thick was used to stop the backscattered protons

RI
and attenuate the low energy X-rays. Measuring times of a selected spot varied in the
range of 600 - 3600 s. The x-ray peak intensities and the concentrations were calculated

SC
by off-line GUPIX program package [22]. The PIXE analysis depth is matrix and element
dependent and can be few tens of micrometers, which is less than the average thickness
of the glaze. It is therefore reasonable to suppose that PIXE interrogates the glaze only

NU
and the analysis is not influenced by the ceramic body. We have calculated that the
penetration depth of protons of 2,2 MeV energy in glaze by SRIM code was about 40 μm,
which is a strong argument confirming that the colourants detected relate solely to the
MA
glaze.
Moreover, Raman spectra of representative glazes were obtained using the same settings
described for the clay analyses. The Raman analysis enhanced our understanding of the
colourants present on the glazes and also provided with an estimation of the firing
D

temperature of the glazes.

TE

Finally, micro-samples were taken from 25 representative samples, in order to examine

the cross-sections. The selection was made so that all different classes, periods and clay
qualities were represented. Cross-sections of the micro-samples were embedded in resin
P

and polished, providing a clear image of the interface between the ceramic body and the
glaze. The microscopic and chemical characteristics of the interface were studied using
CE

SEM/EDS.

3. Ceramic body
AC

3.1. Microscopic characteristics

Polished surfaces of the ceramic body were examined by FOM in order to differentiate
between the different fabrics present among the assemblage. The following
characteristics were taken into consideration: colour (based on the Munsell Soil Colour
Chart [23]), porosity, inclusions (size, shape and frequency) and voids (size and
frequency). Inclusions were defined as ‘coarse’, when larger than 1.0 mm in diameter,
‘medium’, when between 1.0 and 0.5 mm, and ‘fine’, when smaller than 0.5 mm. The
amount of inclusions was defined as ‘abundant’ when it covered more than 20% of the
surface of the sample, ‘regular’ when it covered between 10 to 20%, and ‘low’, when it
covered less than 10%. The size and the amount of the inclusions was estimated by eye.
The microscopic analysis led to the identification of three basic whiteware and four
redware fabrics (Figures 1 and 2). The main characteristics of each fabric are presented in
Table 3.
It should be highlighted that the three major redware groups fit well with the description
of the major fabrics of Corinthian glazed pottery of the same period, provided by Sanders
[2]. More specifically, ‘Fine’, ‘Coarse’ and ‘Pink’ fabrics are the equivalent of Fabrics
D, B and A respectively therein.
 ACCEPTED MANUSCRIPT

Table 3. Microscopic characteristics of the different fabrics present among the samples of the
assemblage

Figure 1. Representative examples of whiteware fabrics (FOM images, magnification x50)

PT
Figure 2. Representative examples of redware fabrics (FOM images, magnification x50)

RI
3.2. Chemical composition
The concentration of the major elements of the ceramic bodies was determined by

SC
SEM/EDS (Table 4). The compositional traits of the clay are in agreement with the
microscopic categorization of the fabrics. Figure 3 shows the silica versus alumina
concentration of the clay bodies, which is indicative of the clay quality of each fabric.

NU
Taking also into consideration the calcium content, the following grouping can be made:
o Group I: The vast majority of the whiteware samples belong to this well-defined
group, including all 15 samples of whiteware ‘Fine’ and ‘Pink’ Fabrics. Samples
of Group I are non-calcareous (~1.8 wt%) with high alumina content (~27 wt. %).
MA
They also have relatively high concentrations of titanium (~1.3 wt%) and iron
(~2.6 wt%).
Three samples have a higher calcium content of approximately 5 wt%; this
increase should be attributed to the presence of an outer layer of very fine
D

calcareous whiteware. The slip was applied in several whiteware samples, in

TE

some cases in thick layers, in order to improve the visual characteristics of the
ceramic objects.
o Group II: The two whiteware samples with ‘Coarse’ fabric have a distinctly
P

different composition. The very high silica content (~80 wt %) and the low
CE

alumina content (~ 8.9 wt %) suggest that these two samples are in fact
stonepaste. Stonepaste is a siliceous ceramic body which was used to make all the
important types of Islamic ceramic. According to archaeological data, stonepaste
was in use since at least 1179 AD [24]. Archaeological recipes for stonepaste
AC

suggest a mixture comprising about ten parts quartz, one part frit-glass, and one
part of a fine white clay [24]. The SEM images clearly show that the samples are
indeed stonepaste (Figure 4). The surface of these samples is mostly covered by
large quartz grains which are bonded together by interparticle glass [25]. It should
be highlighted that these two samples are the only whitewares with a blue-
coloured glaze.
Table 4. Composition of major elements of the ceramic bodies estimated by SEM/EDS. The
mean group values are given as wt % (normalized to 100%); in parentheses the calculated
standard deviations

o Group A: The 17 redware samples of ‘Fine’ fabric form a tight cluster with low
calcium (~6.8 wt%) and high aluminium concentration (~20.1 wt%). They also
have a relatively high alkali content (~1.8 wt% Na2O and ~4.3 wt% K2O).
o Group B: The redware samples of ‘Coarse’ fabric along with four of the samples
of ‘Brown coarse’ fabric form a wide cluster with an aluminium concentration of
approximately 16.8 wt% and a calcium content varying from 6.5 to 14.5 wt%.
 ACCEPTED MANUSCRIPT

o Group C: The 14 redware samples of ‘Pink’ fabric form a separate compositional

group, with an aluminium concentration of approximately 14.4 wt%, low silica
concentration (bellow 53 wt%) and very high calcium compositions (varying
widely from 17 to 31 wt%).
o Group D: Three samples of ‘Brown coarse’ fabric form a distinct group. These are

PT
the only non-calcareous redware samples, with an aluminium content of ~18.4
wt%, and a high silica (~61 wt%) and iron content (~10.3 wt%).

RI
Figure 3. Silica versus alumina concentration of the ceramic bodies

Figure 4. Representative SEM image of a Group II sample. The quartz grains are bonded together

SC
by interparticle glass, suggesting a stonepaste body

Principal Components Analysis (PCA) was conducted based on the SEM/EDS, taking

NU
into consideration all analysed oxides (Figure 5). Additionally, PCA was conducted for
the PGAA data and more specifically the following major and trace elements: H, B, Na,
Mg, Al, Si, Cl, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Nd, Sm, Gd. The related figure is presented
MA
as inset in Figure 5.
Unfortunately, no sample of Group D was analysed by PGAA due to restrictions in the
measuring time. However, the results of both analyses are in good agreement and
corroborate the division in the above mentioned major compositional groups for both the
D

whiteware and redware fabrics.

TE

Figure 5. PCA of the ceramic bodies based on SEM data, along with the related vectors plot. (Top
inset plot: PCA of the ceramic bodies based on PGAA)
P

According to the preliminary mineralogical analysis of the clay bodies based on the
Raman analysis of individual inclusions, quartz, haematite and magnetite is present in all
CE

fabrics, though the amount of inclusions varies significantly. Anatase is also present in all
fabrics, with the exception of Group B. Additionally, olivine and albite were identified
only in samples of Group II and Group B respectively. A more detailed mineralogical
AC

analysis using SEM/EDS is currently under way, and will hopefully assist in a more
detailed description of the fabrics.
Identifying the likely sources of the clay is problematic given that glazed ceramics were a
luxury product during this period, produced in a few large production sites and exported
in large numbers throughout the Mediterranean. Therefore, it is difficult to reliably
identify the excavated pottery as local, unless a ceramic workshop is found. In an early
attempt to discuss the major clay sources throughout Greece, Jones concluded that there
was continuity in the selection of clay sources in Corinth, especially for the production of
fine, decorated pottery [26]. He also suggested the parallel use of several clays in the
vicinity of Corinth. However, the Corinthian clay sources used during this period have
not yet been clearly described chemically.
Based on the available archaeological and compositional data, whitewares are considered
an import of eastern origin [7]. Glazed White Wares in particular are attributed to
Constantinople. The two stonepaste samples are also considered imports from the Middle
East [25]. As far as redwares are concerned, Group C is most likely imported, because of
its specific use for particular classes and because its presence is limited to a particular
 ACCEPTED MANUSCRIPT

period, as will be further discussed in Section 5. Group B may well correspond to the
local production, based on its continuous presence through time. Finally, Group A is the
only group that has been reliably identified as an import from Chalkis, based on both its
major and trace composition. Chalkis, in nearby Euboea, was the port of Thebes and a
major production site for luxury glazed ceramics during the 12th c. AD [5]. Further

PT
compositional and statistical analyses are under way and will hopefully provide more
detailed results regarding the provenance of each fabric identified in Corinth.

RI
4. Glaze
The analyzed glazes are generally well preserved, with an average thickness between 60-
100 μm. The few examples of wider glaze layers, in some cases reaching 200 μm are

SC
found exclusively in rims or highly curved fragments. The difference in the glaze layer
can therefore be attributed to deficiencies caused by the glaze application technique.
4.1. Colourants

NU
The colourant used for each colour hue was determined by the PIXE chemical
composition and the identification of individual inclusions by Raman spectroscopy. The
two blue alkali glazes were produced by adding trace quantities of copper oxides. Green
MA
hues on the lead glazes were produced by copper oxides in highly varying concentrations
(0.5-5.5 wt %). The use of copper oxides for green hues has also been documented for the
contemporaneous green glazes recovered at Englianos [12]. The use of copper as a
colorant in glasses and glazes was extensive throughout history and has been well
D

documented. Under oxidizing conditions, the addition of copper ions (Cu2+) can give a
TE

blue or green colour to the glass, depending on the glass modifiers present. Generally,
copper oxide will produce green along with lead oxide and turquoise blue along with
sodium or potassium oxide [27]. Orange, and brown hues were produced by adding iron
P

oxides; however, in a few cases of Green and Brown Painted samples the dark brown
decorative lines were produced by adding manganese. Finally, the majority of yellow
CE

samples show slightly elevated iron concentration. However, in two early 14th c. sherds
inclusions of Naples’ Yellow (lead antimonate) were identified based on their
characteristic Raman bands at 141, 336, 453 and 505 cm-1 [28]. It is interesting to note
AC

that the 12th-13th c. yellow glazes from Englianos were also coloured by the addition of
iron oxides, but in some later sherds of the 15th-16th c. the colourant was a recently
identified lead-tin-antimony oxide, with a distinctive Raman spectrum [12, 29]. Two
likely explanations can be provided for the diversity noted on the selection of yellow
colourants in the later samples of Corinth and Englianos: (a) the two sites belong to
slightly different periods, therefore there could have been a change in the yellow
colourants used through time; or (b) the different workshops could be following specific
recipes which indicated the selective use of a yellow colourant.

4.2. Glaze technique

The chemical analysis of the glazes shows that the vast majority belong to the high-lead
type, though the Pb:Si ratio demonstrates a significant variability which will be further
discussed in Section 5 (Table A3 of the Appendix). The only exceptions are the two blue
glazes which are alkali glazes. The SiO2 concentration varies from 63 to 73 wt% and
Na2O concentration is approximately 20 wt %. Al2O3, K2O, MgO and CaO present in
smaller concentrations. It should be highlighted that these two samples are made of
 ACCEPTED MANUSCRIPT

Group II clay, therefore forming an overall distinct category of pottery that used both
different raw materials and a different production technology.
High-lead glazes can be produced by a number of different methods. The most common
production methods are: (a) use of a lead compound by itself, typically as a suspension in
water; and (b) use of a suspension of a lead compound with silica [30]. In the former

PT
case, the lead diffuses into the ceramic body during firing, reacting with it to form a
glaze. Therefore, if the lead oxide content is subtracted from the glaze composition which
is then re-normalized to l00%, the resulting composition should be close to that of the

RI
body. By contrast, when a lead-silica mixture is used, the lead and silica react together
during firing to form the glaze, with subsequent diffusion of the body components. In this
case, the glaze composition, after subtraction of the lead oxide content and re-

SC
normalizing to l00%, will depend on the relative diffusion rates of the different body
components. Therefore, it will not normally match the body composition [31].
Walton and Tite [32] suggest the use of bi-plots of the concentration of each major oxide

NU
in the ceramic body versus the concentration in the glaze, after extracting the colourants
and re-normalizing. When the data points fall on the unity line, the glaze and body
compositions are essentially the same and, therefore the glaze was produced by the
MA
application of lead oxide by itself. The data plotted in Figure 6 for silica, calcium and
iron oxide, suggest that all of the whiteware glazes were produced by the application of
lead oxide by itself, whereas all redware glazes were produced using a lead-silica
mixture.
D
TE

Figure 6. Bi-plots of SiO2, CaO and FeO concentration of the clay body and of the glaze (after
subtraction of the lead concentration and re-normalization)

4.3. Application technique

P

A further variation of the lead glaze production method is that the lead glaze can be
CE

applied either on unfired or on biscuit-fired ceramic bodies. Distinguishing between the

two application techniques is possible by examining the clay-glaze interface [32, 33].
During firing, elements such as Al, Fe, K, Ca and Mg diffuse from the body into the
AC

glaze and, as the concentration of these elements in the glaze increases, crystals of
potassium-lead-aluminium-silicate are formed in the clay-glaze interface. When the glaze
suspension is applied to an unfired body, the diffusion from the body into the glaze is
significantly greater. As a result, a significantly higher concentration of potassium-lead-
aluminium-silicate crystals are present in the body-glaze interface for an unfired body as
opposed to a biscuit-fired body (Figure 7).
Moreover, the decreased diffusion evident in the case of a biscuit-fired body also results
in differences in the compositional profile (Figure 8). Elements that are not originally
present in the glaze, such as K, Al, Fe and Mg, have low concentrations in a tight zone
close to the clay-glaze interface, after which point their concentration is almost zeroed.
At the same time, lead concentration is almost stable throughout the glaze.
Based on the examined cross-sections it is suggested that the glaze on all whiteware
samples was applied on unfired bodies. By contrary, a number of biscuit-fired examples
were identified among redware sherds.

Figure 7. SEM photographs of body-glaze interfaces for high lead glazes showing (left) the high
concentration of potassium-lead-aluminium-silicate crystals, characteristic of an unfired body; and
 ACCEPTED MANUSCRIPT

(right) the absence of potassium-lead-aluminum-silicate crystals, characteristic of a biscuit-fired

body

Figure 8. Compositional profiles across glaze sections produced by applying glaze on an unfired body
(dashed line) and by applying glaze on a biscuit-fired body (solid line) (0 μm: clay-glaze interface)

PT
4.4. Firing temperature
Colomban et al. have previously discussed the value of using the Raman spectra of
archaeological glasses and glazes for estimating the firing temperature, by calculating the

RI
polymerization index [34, 35, 36]. The two main Raman bands of glass correspond to the
stretching of Si-O- non-bridging bonds at 960 and 1100 cm-1 in Figure 9, and to the
mixed stretching-bending vibration, ν(Si-O-Si), of Si–O-Si bridging bonds measured at

SC
480 and 570 cm-1 in Figure 9. The intensity ratio of these two bands determines the glass
polymerization and is a function of two factors: (a) the bulk composition; and (b) the

NU
firing temperature. Colomban [34] thoroughly describes a specific statistical protocol for
the calculation of the polymerization index, Ip, (i.e. the A500/A1000 ratio). The study of a
variety of different archaeological glasses and glazes led to the identification of six
different ‘families’ according to their Ip value, each characterised by a specific
MA
composition and firing temperature [34, 35, 36].

Figure 9. Characteristic Raman spectra of a blue alkali glaze and a lead glaze
D

Figure 10. Bi-plot of the Polymerization Index (Ip) versus the PbO:SiO 2 ratio of lead glazes and alkali
glazes
TE

Following the same protocol we can conclude that all of the lead glazes in the assemblage
have an Ip value lower than 0.3 (Figure 10). The estimated Ip values fit well with the
P

expected values for high lead glazes and suggest a low firing temperature of
CE

approximately 750 ºC. The SEM examination of the cross sections also suggests a low
firing temperature, based on the width of the interface zone and the size of the potassium-
lead-aluminium-silicate crystals formed [33]. One lead glaze with a high Ip value (0.57)
AC

has a highly weathered surface; therefore the noted differentiation of the polymerization
index is attributed to the corrosion effects on the glass matrix. Finally, the two blue alkali
glazes have distinctively higher Ip values (~0.68), which are however lower than the
documented Ip values for alkali glazes fired at approximately 1000 ºC (0.8-1.1). The
noted difference is most likely not related to differences in the firing temperature, but due
to compositional variations, since the two samples under study have higher Na2O values
than most alkali glazes.

5. Changes in redware production in the beginning of the 13th c. AD

The redware samples of the assemblage demonstrate a high degree of variation in both
their clay and glaze composition. An examination of the correlation between clay fabrics
and type of decoration reveals that clay Groups A and B were used in a highly specialized
production and associated primarily to high-quality luxury items, as opposed to Group C
(Figure 11a).
More specifically, the unglazed and glazed vessels with no other decoration were
produced using clay Groups B and D. Sgraffito Wares were produced mainly using
Groups A and B. In fact, Group A prevails in all sgraffito classes, with the exception of
 ACCEPTED MANUSCRIPT

the Late Sgraffito Ware, indicating a change in the clay sources used during the later
periods. Group C was only used for Fine Sgraffito Ware, indicating the selective use of
raw materials for each one of these luxury types of ware. ‘Painted’ wares were produced
by all three major fabrics in approximately equal amounts. However, specific fabrics are
associated to specific decorative techniques. Slip Painted (Dark on Light) and Measles

PT
Ware were produced exclusively by Group C fabrics. Green and Brown Painted Ware
was mostly produced by Groups A and C fabrics. Slip Painted Ware was produced by
Groups A or B fabrics.

RI
Furthermore, the selection of clay fabrics shows a clear shift through time (Figure 11b).
Group B was in use continually from the 10th to the 12th c. AD. During the 13th c. its use

SC
increased significantly, leading to the exclusive use of this fabric during the 14th c. AD.
Group A was present from late 11th c. to the early 13th c., when its use stopped. In the 12th
c., almost 40% of all samples were produced in Group A fabrics. Finally, the use of

NU
Group C is restricted to the 12th c., which it dominated. Over 50% of the 12th c. samples
were produced in Group C fabrics.
During the 13th c. political fragmentation occurred and significantly altered all aspects of
MA
life in the rural areas of the Byzantine Empire. As far as ceramic production is concerned,
the predominance of a few specialized workshops seized and trade among cities was
minimized. Smaller regional workshops covered needs of local markets, using local raw
materials, often of lower quality.
D
TE

Figure 11. Correlation of redware clay fabrics with type of decoration (left) and chronology (right)

The composition of the redware glazes seem to be correlated with the production date
P

(Figure 12a). Glazes produced in the 13th c. AD show smaller variation of the PbO:SiO2
ratio (between 1.3 and 2.1) and an overall more homogenous composition. On the
CE

contrary, the PbO:SiO2 ratio of 11th and 12th c. glazes varies greatly (between 0.9 and
2.4). The earlier glazes also demonstrate a much higher variability in the total
concentration of alkalis and alkaline earths.
AC

By contrast glaze composition is not correlated with the type of decoration (using
Morgan’s [1] division into three categories) (Figure 12a). Therefore, the changes in the
glaze composition at the beginning of the 13th c. AD should not be attributed to the
selective production of different styles of pottery, but reveal a more general technological
trend.
The difference in the glaze composition can be attributed to the introduction of
technological innovations in the glazing process. This was one of a number of such
innovations introduced in the beginning of the 13th c. AD, which included the use of
tripods during the heating of the glaze, allowing for better control of the firing conditions
[5, 37, 38]. Given the high volatility of lead, the improved control of firing conditions
may account for both the increased PbO:SiO2 ratio and the increase in homogeneity.

Figure 12. Correlation of glaze composition with type of decoration (left) and chronology (right)
 ACCEPTED MANUSCRIPT

6. Concluding remarks
The present work focuses on the application of different analytical techniques and
statistical tools in order to characterize fully the analytical and technological properties of
an assemblage of Byzantine glazed ceramics from Corinth. The study of the ceramic
body, glaze and clay-glaze interface was achieved through the combined use of analytical

PT
techniques (SEM/EDS, PGAA, milli-PIXE) with Raman spectroscopy. In addition,
statistical tools (PCA) and methodological approaches (e.g. Hurst and Freestone’s
methodology for high-lead glaze discrimination, calculation of Ip) were used in order to

RI
examine parameters of the manufacturing process.
The analysis of whiteware revealed a uniform production. The vast majority of
whitewares were made of the same raw materials (clay Group I), whose provenance,

SC
however, has not yet been identified. A lead compound by itself was used for the
production of the high-lead glaze, which was applied on unfired ceramic bodies. Only
two whiteware samples differ from the norm, forming a distinct group (Group II)

NU
produced by the application of alkali glaze on a stonepaste body.
By contrast, the redware samples show a high variation in terms of fabrics, glaze
composition and manufacturing techniques. Three main fabrics are identified: Group A
MA
originates in Chalkis, Group C is likely another import of unidentified origin, whereas
Group B most probably corresponds to the local Corinthian production. Each fabric
corresponds to specific classes of pottery. In the 13th c. AD, the few specialized
workshops seized and the trade among cities was minimized. In their place, smaller
D

regional workshops that hitherto dominated production were replaced by many smaller
TE

regional workshops, which used local raw materials of lower quality in order to cover
local demands.
The glazes of the redware samples were produced using a suspension of lead with silica,
P

which was applied either on unfired or biscuit-fired bodies. The firing temperature of the
glaze is estimated at approximately 750 ◦C. The introduction of technological innovations
CE

in the glazing process, which occurred in the early 13th c. AD, resulted in a better control
of firing conditions and consequently, in more homogenous glazes of improved quality,
despite the fragmentation of production. The different source of yellow colourants in the
AC

later samples of Korinthos and Englianos can serve as a further indication of localized
production.
Overall, the present study reveals a sophisticated network of glazed pottery production
and distribution during the Middle Byzantine and early Late Byzantine periods.
Furthermore, it highlights the significance of the analysis of Byzantine glazed pottery
from a technological point of view, which, in addition to ‘chaîne opératoire’ studies, shed
light to the ceramic production of this period.

Acknowledgements
EP gratefully acknowledges financial support from the PELOPS Scholarship Program of
the University of the Peloponnese. NZ acknowledges a grant provided by the
CHARISMA-FP7 (contract No. 228330) covering transportation, accommodation and
experiments at FIXLAB facilities (PGAA and PIXE) of the Budapest Neutron Centre, for
the performance of PGAA and PIXE measurements. The authors acknowledge permits
 ACCEPTED MANUSCRIPT

granted by the excavator Guy Sanders and the Greek Ministry of Culture and Sports to
enable the study of the assemblage.

References
[1] C. Morgan, Corinth XI. The Byzantine pottery, Mass, Cambridge, 1942.

PT
[2] G.D.R. Sanders, Byzantine glazed pottery at Corinth to c.1125, DPhil thesis,
University of Birmingham, 1995.

RI
[3] G.D.R. Sanders, New relative and absolute chronologies for 9th to 13th century Glazed
wares at Corinth: Methodology and social conclusions, Byzanz als Raum. Zu Methoden

SC
und Inhalten der historischen Geographie des östlischen Mittelmeerraumes im Mittelalter,
Vienna, 2000, pp. 153-173.
[4] G.D.R. Sanders, Recent Developments in the Chronology of Byzantine Corinth, in: C.

NU
K. Williams, N. Bookidis (Eds.), Corinth XX, The Centenary, 1896-1996, 2003, pp. 381-
394.
[5] S.Y. Waksman, N.D. Kontogiannis, S.S. Skartsis, G. Vaxevanis, The main ‘Middle
MA
Byzantine production’ and pottery manufacture in Thebes and Chalkis, The annual of the
British School at Athens, 109 (2014) 379-422.
[6] S.Y. Waksman, I. Teslenko, “Novy Svet Ware”, an exceptional cargo of glazed wares
D

from a 13th-century shipwreck near Sudak (Crimea). Morphological typology and

laboratory investigations, International Journal of Nautical Archaeology, 39, 2 (2010)
TE

357-375.
[7] S.Y. Waksman, A. Bouquillon, N. Cantin, I. Katona, The first Byzantine “Glazed
P

White Wares” in the Early Medieval technological context, in: S.Y. Waksman (Eds.),
CE

Archaeometric and archaeological approaches to ceramics (papers presented at EMAC

2005), BAR International Series 1691, Oxford, 2007, pp. 129-135.
[8] S.Y. Waksman, E.J. Stern, I. Segal, N. Porat, J. Yellin, Elemental and petrographic
AC

analyses of local and imported ceramics from Crusader Acre, Atiqot, 59 (2008) 157-190.
[9] S.Y. Waksman, Ceramics of the “Serçe Limanı type” and Fatimid pottery production
in Beirut, Levant, 43, 2 (2011) 201-212.
[10] A.C. Charalambous, A.J. Sakalis, N.A. Kantiranis, L.C. Papadopoulou, N.C.
Tsirligiannis, J.A. Stratis, Cypriot Byzantine glazed pottery: A study of the Paphos
workshops, Archaeometry, 52, 4 (2010) 628-643.
[11] P. Armstrong, H. Hatcher, M.S. Tite, Changes in Byzantine glazing technology from
the ninth to thirteenth centuries, La Ceramique Medievale en Mediterranee, Actes du 6
Congress, Aix-en-Provence, 6 (1997) 225-229.
[12] J.L. Davis, S.R. Stocker, The Medieval deposit from the Northeast Gateway at the
Palace of Nestor, Hesperia: The Journal of the American School of Classical Studies at
Athens, 82, 4 (2013) 673-731.
[13] D. Sokaras, A.G. Karydas, A. Oikonomou, N. Zacharias, K. Beltsios, V. Kantarelou,
Combined elemental analysis of ancient glass beads by means of ion beam, portable
 ACCEPTED MANUSCRIPT

XRF, and EPMA techniques, Analytical and Bioanalytical Chemistry, 395 (2009) 2199-
2209.
[14] D. Bersani, P.P. Lottici, S. Virgenti, A. Sodo, G. Malvestuto, A. Botti, E. Salvioli-
Mariani, M. Tribaudino, F. Ospitali, M. Catarsi, Multi-technique investigation of
archaeological pottery from Parma (Italy), Journal of Raman Spectroscopy, 41 (2010)

PT
1556-1561.
[15] L. Szentmiklósi, T. Belgya, Zs. Révay Z. Kis, Upgrade of the prompt gamma

RI
activation analysis and the neutron-induced prompt gamma spectroscopy facilities at the
Budapest research reactor, Journal of Radioanalytical and Nuclear Chemistry, 286 (2010)
501-505.

SC
[16] Zs. Révay, Determining elemental composition using prompt gamma activation
analysis, Analytical Chemistry, 81 (2009) 6851-6859.

NU
[17] V. Szilágyi, J. Gyarmati, M. Tóth, H. Taubald, M. Balla, Zs. Kasztovszky, Gy.
Szakmány, Petro-mineralogy and geochemistry as tools of provenance analysis on
archaeological pottery: Study of Inka Period ceramics from Paria, Bolivia, Journal of
MA
South American Earth Sciences, 36 (2012) 1-17.
[18] A. Oikonomou, K. Beltsios, N. Zacharias, Analytical and technological study of blue
glass from Thebes, Greece: An overall assessment, in: D. Ignatiadou, A. Antonaras (Eds),
Proceedings of the AIHV 18, Thessaloniki, 2012, pp. 75-80.
D

[19] I. Gyódi, I. Demeter, I., K. Hollós-Nagy, I. Kovács, Z. Szőkefalvi-Nagy, External-

TE

beam PIXE analysis of small sculptures, Nuclear Instruments and Methods in Physics
Research Section B: Beam Interactions with Materials and Atoms, 150, 1-4 (1999) 605-
610.
P

[20] B. Constantinescu, D. Cristea-Stan, I. Kovács, Z. Szőkefalvi-Nagy, External milli-

CE

beam PIXE analysis of the mineral pigments of glazed Iznik (Turkey) ceramics,
Periodico di Mineralogia, 83, 2 (2014) 159-169.
AC

[21] J. F. Ziegler, J. P. Biersack, M. D. Ziegler, The stopping and range of ions in matter,
SRIM (2008).
[22] J.L. Campbell, T.L. Hopman, J.A. Maxwell, Z. Nejedly, The Guelph PIXE software
package III: alternative proton database, Nuclear Instruments and Methods B, 170 (2000)
193-204.
[23] Munsell Soil Colour Charts (2000).
[24] R.B. Mason, M.S. Tite, The beginnings of Islamic stonepaste technology,
Archaeometry, 36, 1 (1994) 77-91.
[25] M.S. Tite, S. Wolf, R.B. Mason, The technological development of stonepaste
ceramics from the Islamic Middle East, Journal of Archaeological Science, 38 (2011)
570-580.
[26] R.E. Jones, Greek and Cypriot pottery: A review of scientific studies, Athens:
British School at Athens (1986).
 ACCEPTED MANUSCRIPT

[27] S. Davison, Conservation and restoration of glass (2nd edition), Butterworth-

Heineman, London (2003).
[28] Ph. Colomban, G. Sagon, X. Faurel, Differentiation of antique ceramics from the
Raman spectra of their coloured glazes and paintings, Journal of Raman Spectroscopy, 32
(2001) 351-360.

PT
[29] C. Pelosi, G. Agresti, U. Santamaria, E. Mattei, Artificial yellow pigments:
Production and characterization through spectroscopic methods of analysis, E-

RI
Preservation Science (2010) Morana RTD.
[30] M.S. Tite, I. Freestone, R. Mason, J. Molera, M. Vendrell-Saz, N. Wood, N. Lead

SC
glazes in antiquity – Methods of production and reasons for use, Archaeometry, 40, 2
(1998) 241-260.

NU
[31] D. Hurst, I. Freestone, Lead glazing technique from a medieval kiln site at Hanley
Swan, Worcestershire, Medieval Ceramics, 20 (1996) 13-18.
[32] M.S. Walton, M.S. Tite, Production technology of Roman lead-glazed pottery and its
MA
continuance into late antiquity, Archaeometry, 52, 5 (2010) 733-759.
[33] J. Molera, T. Pradell, N. Salvado, M. Vendrell-Saz, Interactions between clay bodies
and lead glazes, Journal of the American Ceramic Society, 84, 5 (2001) 1120-28.
D

[34] Ph. Colomban, Polymerization degree and Raman identification of ancient glasses
TE

used for jewelry, ceramic enamels and mosaics, Journal of Non-Crystalline Solids, 323
(2003) 180-187.
P

[35] Ph. Colomban, A. Tournié, L. Bellot-Gurlet, Raman identification of glassy silicates

used in ceramics, glass and jewelry: a tentative differentiation guide, Journal of Raman
CE

Spectroscopy, 37 (2006) 841-852.

[36] Ph. Colomban, M.-P. Etcheverry, M. Asquier, M. Bounichou, A. Tournié, Raman
AC

identification of ancient stained glasses and their degree of deterioration, Journal of

Raman Spectroscopy, 37 (2006) 614-626.
[37] M. Giannopoulou, Introduction to lead glazes in antiquity. Glazing technology in
traditional pottery, in: P. Themelis (Eds.) Glass-From Antiquity until today, Society of
Messenian Archaeological Studies, Athens (2002) (in Greek).
[38] D. Papanikola-Bakirtzis, Heating tripods of Byzantine and late-Byzantine vessels, in
Amitos – Special Volume in the Honour of Prof. Manolis Andronikos, Thessaloniki
(1997) (in Greek).
 ACCEPTED MANUSCRIPT

Table A1. SEM/EDS composition of major elements of all analysed ceramic bodies

Sample Na2O MgO Al2O3 SiO2 K2O CaO TiO2 FeO

BYZ51 0.85 2.36 17.45 60.94 2.92 1.54 1.03 12.92
BYZ52 0.67 2.53 17.01 54.17 2.51 11.87 1.02 10.22
BYZ53 1.02 2.09 19.04 62.26 2.92 1.64 0.91 10.10

PT
BYZ55 0.32 0.60 27.35 66.28 0.77 0.58 1.35 2.75
BYZ56 0.18 0.62 28.29 66.01 0.93 0.52 1.22 2.23
BYZ57 0.61 1.66 25.77 60.32 1.30 6.25 0.94 3.14

RI
BYZ58 0.19 0.48 28.88 65.74 0.69 0.52 1.31 2.18
BYZ59 0.23 0.51 26.64 68.15 0.89 0.45 1.27 1.87
BYZ60 0.57 1.00 25.49 64.27 1.16 4.22 1.47 1.84

SC
BYZ61 0.23 0.67 29.02 65.12 0.73 0.62 1.08 2.53
BYZ62 0.14 0.59 25.39 66.16 1.05 0.87 1.42 4.39
BYZ63 0.72 1.56 20.92 63.26 3.34 5.92 1.48 2.81

NU
BYZ64 0.21 0.53 25.67 68.54 0.92 0.53 1.57 2.03
BYZ66 2.62 1.22 8.66 81.25 2.20 2.47 0.48 1.11
BYZ67 0.18 0.59 28.89 65.15 1.00 0.73 1.27 2.20
BYZ68 0.00 1.14 23.36 64.60 2.07 4.51 1.05 3.28
MA
BYZ69 0.28 0.80 28.04 65.04 1.27 0.85 1.20 2.52
BYZ70 0.22 0.62 28.01 66.17 0.86 0.60 1.27 2.25
BYZ71 0.39 0.71 33.17 59.73 1.83 0.50 0.96 2.70
BYZ72 1.68 2.39 20.71 55.73 4.19 6.52 0.90 7.88
D

BYZ73 1.69 2.70 20.25 56.80 3.64 6.95 0.68 7.28

BYZ74 0.67 3.17 14.21 44.99 3.06 23.83 0.81 9.26
TE

BYZ75 1.13 3.85 14.36 43.62 1.40 26.71 0.72 8.21

BYZ76 1.02 3.39 15.44 49.31 3.00 19.84 0.78 7.22
BYZ77 1.60 2.66 19.94 55.61 4.05 6.54 0.82 8.78
P

BYZ78 1.48 2.76 19.27 55.95 4.09 6.50 0.95 9.00

BYZ79 0.96 3.18 15.41 49.37 3.20 18.72 0.91 8.24
CE

BYZ80 1.28 2.57 19.19 55.16 4.32 7.49 0.86 9.12

BYZ81 1.16 2.75 16.59 58.63 4.20 7.16 0.69 8.82
BYZ82 1.20 3.67 17.53 56.92 4.47 6.54 0.83 8.84
AC

BYZ83 1.43 3.07 16.48 56.95 3.97 8.37 0.74 9.00

BYZ84 1.02 3.48 17.45 58.37 4.00 6.53 0.94 8.20
BYZ85 0.84 5.80 15.71 54.37 3.20 11.09 0.64 8.35
BYZ86 0.67 2.39 16.94 59.44 2.81 9.04 0.86 7.86
BYZ87 0.96 2.38 15.90 56.51 3.05 12.81 0.82 7.57
BYZ88 1.88 2.59 20.48 54.24 4.18 7.02 0.81 8.79
BYZ89 1.36 2.54 20.91 54.72 4.96 6.76 0.87 7.87
BYZ90 1.60 2.68 21.09 54.52 4.65 7.27 0.83 7.36
BYZ91 1.04 5.26 16.39 49.67 4.03 10.79 0.81 12.01
BYZ92 1.34 5.54 16.75 47.90 3.72 14.91 0.95 8.90
BYZ93 1.35 3.46 18.72 57.48 4.34 5.57 0.74 8.33
BYZ94 1.10 3.04 15.46 52.46 1.87 18.96 0.75 6.35
BYZ95 0.82 3.62 14.95 43.76 2.71 23.26 1.14 9.74
BYZ96 0.95 3.59 14.67 46.67 3.24 23.95 0.58 6.35
BYZ97 2.00 3.04 21.14 53.74 4.92 6.77 0.98 7.41
BYZ98 1.96 0.71 9.12 78.60 2.06 6.25 0.36 0.95
BYZ99 0.57 3.08 16.89 53.54 2.53 16.61 0.65 6.13
BYZ100 0.65 2.56 13.86 45.26 2.75 24.85 0.87 9.22
BYZ101 2.96 2.57 20.77 52.50 4.95 7.55 0.88 7.81
 ACCEPTED MANUSCRIPT

BYZ102 1.91 2.60 19.86 53.78 4.22 7.61 0.93 9.10

BYZ103 0.75 2.80 13.34 42.80 3.37 27.59 1.02 8.32
BYZ104 1.67 2.52 19.81 56.33 4.21 5.97 1.04 8.45
BYZ105 1.46 2.86 20.64 56.47 4.09 5.92 0.73 7.83
BYZ106 1.73 2.57 19.40 55.42 3.99 7.13 0.79 8.97
BYZ107 1.42 2.44 19.79 55.82 4.10 7.12 0.85 8.46

PT
BYZ108 1.61 2.79 17.20 52.83 3.60 13.32 0.89 7.76
BYZ109 1.78 2.65 18.44 55.46 3.59 9.93 0.89 7.26
BYZ110 1.54 2.54 18.44 56.50 3.82 8.54 0.79 7.83

RI
BYZ111 2.23 2.93 19.11 55.88 4.02 7.14 0.80 7.88
BYZ112 1.91 2.94 19.81 52.76 4.29 9.33 0.76 8.19
BYZ113 1.09 4.10 18.58 60.50 4.52 2.70 0.65 7.86

SC
BYZ114 2.06 3.18 19.68 55.82 3.74 6.91 0.85 7.77
BYZ115 0.88 3.00 15.62 50.87 3.67 17.67 0.78 7.50
BYZ116 1.03 2.84 14.13 45.00 3.05 24.51 0.90 8.54

NU
BYZ117 0.87 2.76 13.00 42.33 2.97 30.35 0.77 6.97
BYZ118 0.58 3.09 12.44 40.86 3.11 31.10 0.57 8.24
BYZ119 0.78 4.72 15.44 55.14 3.77 9.02 0.96 10.17
BYZ120 0.76 5.17 16.43 52.95 3.95 11.76 0.77 8.20
MA
BYZ121 0.84 5.29 17.74 54.61 3.53 8.82 0.79 8.37
BYZ149 0.73 2.88 16.42 58.25 2.88 9.90 0.77 8.19
BYZ150 0.48 2.52 16.46 58.89 2.94 10.46 0.79 7.46
BYZ151 0.29 2.68 12.65 61.10 3.21 10.91 0.90 8.26
D

BYZ152 0.81 2.53 17.06 59.74 2.99 7.48 0.92 8.47

BYZ157 0.54 2.78 15.96 53.59 2.62 16.48 1.05 6.97
P TE
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE
P
CE
AC
52 66 69 70 72 81 83 87 89 94 97 98 101 107 112 117 120
0.69 1.95 0.21 0.00 1.28 1.45 1.31 0.62 1.38 0.87 1.36 0.92 1.51 1.40 1.34 0.87 0.91
0.00 0.83 0.79 0.00 2.94 2.17 2.62 2.03 2.40 2.98 3.09 0.00 2.62 2.64 3.08 3.40 4.35
AC
15.93 2.70 24.56 22.67 19.43 14.88
CE 14.44 14.40 19.20 12.95 20.00 5.07 19.38 18.60 21.25 14.56 15.42
65.42 91.30 69.59 72.43 58.82 67.16 66.97 63.88 59.47 56.52 56.71 90.53 58.66 59.98 56.15 46.11 57.97
2.25 1.36 0.89 0.70 3.83 3.10
P 3.07 2.24 3.96 1.99 4.23 1.10 3.98 3.96 4.07 3.05 3.73
7.18 1.36 0.46 0.48 5.25 4.44 4.81 9.02 5.32 17.06 5.67 1.79 5.66 4.88 5.39 24.14 8.80
TE
0.83 0.22 1.14 1.20 0.88 0.81 0.72 0.81 0.86 0.74
D 0.89 0.21 0.86 0.90 0.89 0.72 0.85
7.70 0.29 2.35 2.52 7.56 5.99 6.07 7.00 7.41 6.89 8.04 0.39 7.33 7.62 7.82 7.16 7.97
3924 785 2733 4998 4715 3662 5231 1396 4500 2587 2641 7822 4133 1398 9434 1830 3456
MA
69 35 42 36 77 62 67 64 82 73 84 23 80 78 88 659 81
5136 nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd
NU
695 nd nd nd nd nd nd nd nd nd 1767 nd nd nd nd 1351 nd
275 125 115 185 239 469 385 1398 1251 971 1167 143 3020 311 930 542 158
SC
58 3 36 36 13 16 17 20 22 19 11 2 22 15 26 12 21
125 36 174 257 145 184 107 140 167 166 188 33 158 153 186 131 179
RI
elements presented as oxides, in wt% and trace elements as elements, in ppm)

609 nd 92 nd 256 377 329 nd nd 323 233 nd 268 285 205 253 611
ACCEPTED MANUSCRIPT

PT
934 91 83 114 856 383 371 830 895 1108 1091 125 850 1054 871 904 866
35 9 15 19 39 32 28 34 38 33 44 14 36 26 40 31 39
5 1 2 2 5 4 3 4 5 4 5 2 5 5 5 3 4
5 2 2 3 5 4 4 4 6 4 5 1 6 6 5 4 5
Table A2. PGAA composition of major and trace elements of representative ceramic bodies (major
 Na2O
MgO
AC
CE Al2O3
SiO2
P K 2O
CaO
TE
D TiO2
Fe2O3
H
MA
B
P
NU
S
Cl
SC
Sc
V
RI
Cr
ACCEPTED MANUSCRIPT

PT
Mn
Nd
Sm
Gd
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

Table A3. SEM/EDS composition of major elements of all analysed glazes

Sample Colour Na2O MgO Al2O3 SiO2 K2O CaO TiO2 MnO FeO CuO PbO
BYZ51 Dark Brown 0.33 1.20 8.14 35.68 0.76 0.90 0.40 nd 2.94 nd 49.66
BYZ52 Dark Brown 0.42 1.25 8.05 36.74 0.79 4.61 0.39 nd 2.57 nd 45.17
BYZ56 Yellow nd 0.20 9.29 28.98 0.11 0.24 nd nd 0.29 nd 60.91

PT
BYZ56 Light Green 0.27 0.40 9.12 27.82 1.46 0.83 0.54 nd 0.87 0.33 58.35
BYZ59 Green nd 0.45 10.33 30.13 nd 0.31 nd nd nd nd 58.79
BYZ62 Orange nd 0.41 11.34 33.52 0.14 nd 0.26 nd 0.58 nd 53.79

RI
BYZ66 Blue 21.87 4.57 2.22 63.51 1.57 5.93 nd nd 0.34 nd nd
BYZ67 Yellow 0.31 0.34 11.10 32.27 nd 0.17 0.36 nd 0.23 nd 55.23
BYZ67 Brown-Red 1.93 0.45 11.60 29.32 0.68 1.04 0.42 nd 2.06 nd 52.49

SC
BYZ69 Yellow nd 0.29 10.10 32.54 0.18 0.42 nd nd 0.21 nd 56.25
BYZ70 Green nd 0.39 10.18 30.05 0.34 0.35 0.39 nd 0.45 0.54 57.29
BYZ70 Dark Brown nd 0.33 9.14 29.58 0.25 0.18 0.25 nd 0.49 nd 59.79

NU
BYZ71 Light Green nd 0.25 5.24 40.15 0.27 0.41 0.21 nd 0.37 0.24 52.84
BYZ72 Cream nd 0.29 4.08 35.85 0.14 0.24 nd nd 0.20 nd 59.20
BYZ72 Dark Brown nd 0.44 3.97 35.06 0.13 0.70 nd 0.89 0.21 nd 58.61
BYZ73 Cream 0.15 0.25 1.40 37.35 nd 0.29 nd nd nd nd 60.57
MA
BYZ74 Dark Brown 0.18 0.35 1.54 27.98 0.28 1.84 nd nd 0.24 nd 67.60
BYZ74 Cream 0.32 0.69 3.28 28.52 0.63 2.69 nd nd 0.58 nd 63.31
BYZ76 Cream 0.37 0.39 3.32 37.83 1.00 1.15 nd nd 0.46 nd 55.47
BYZ76 Green nd 0.44 1.61 37.97 0.35 1.05 nd nd 0.39 2.09 56.14
D

BYZ76 Dark Brown nd 0.57 2.40 38.74 0.55 0.78 nd 2.31 0.52 nd 54.12
BYZ78 Brown 0.51 0.68 6.87 44.90 0.54 1.58 nd 1.60 2.31 nd 41.01
TE

BYZ79 Green 0.28 0.18 5.21 32.86 1.16 1.02 0.26 nd 0.46 3.07 55.52
BYZ81 Yellow 0.93 1.27 6.84 29.60 0.68 0.34 nd nd 0.48 nd 59.86
BYZ81 Green 0.74 1.08 6.55 27.84 1.58 1.42 0.14 nd 1.17 nd 59.01
P

BYZ83 Yellow 0.36 0.97 5.16 33.44 0.68 1.16 0.16 nd 0.65 nd 57.43
BYZ83 Green 0.38 1.19 6.38 34.45 0.72 1.42 0.22 nd 0.71 1.92 52.60
CE

BYZ85 Brown 0.44 1.22 4.79 30.67 0.51 2.47 nd nd 2.00 nd 57.91
BYZ86 Brown 0.20 0.37 2.91 30.09 0.38 0.76 nd nd 0.42 nd 64.88
BYZ86 Dark Brown 0.17 0.29 1.60 35.96 0.24 0.87 nd nd 1.30 nd 59.57
AC

BYZ87 Yellow 0.24 0.41 3.43 29.32 0.47 0.66 0.14 nd 0.44 nd 64.89
BYZ87 Brown 0.14 0.43 2.49 30.02 0.29 1.29 nd nd 0.56 nd 64.78
BYZ89 Brown 0.35 0.61 4.13 31.17 1.01 1.44 0.15 nd 2.43 nd 58.71
BYZ90 Green 0.24 0.87 4.48 33.28 0.34 1.38 nd nd 0.93 0.96 57.51
BYZ90 Brown 0.32 1.03 4.87 34.26 0.50 1.59 0.14 nd 1.03 1.00 55.28
BYZ93 Yellow 0.65 0.85 5.93 34.55 1.50 1.02 0.30 nd 0.92 nd 54.27
BYZ93 Green 0.66 0.78 5.05 29.93 0.94 0.66 0.39 nd 1.26 4.45 55.88
BYZ93 Brown 0.64 1.26 5.75 31.54 1.08 2.02 0.50 nd 2.66 nd 54.55
BYZ94 Yellow 0.16 0.57 2.39 40.44 0.70 0.38 nd nd 0.52 nd 54.86
BYZ94 Brown 0.14 0.56 3.58 40.13 0.89 0.43 0.17 nd 0.68 nd 53.44
BYZ96 Yellow 0.34 0.28 1.96 31.40 0.72 1.88 nd nd 2.97 nd 60.25
BYZ96 Brown 0.31 0.76 3.15 36.25 0.36 0.65 nd nd 0.36 nd 58.15
BYZ97 Dark Brown 0.24 0.49 2.32 36.14 0.10 1.41 nd 1.41 2.11 nd 55.77
BYZ98 Blue 19.33 2.19 4.24 72.67 0.78 0.79 nd nd nd nd nd
BYZ99 Yellow 0.38 0.84 4.43 33.02 0.58 1.52 nd nd 0.33 nd 58.90
BYZ101 Yellow 0.60 0.41 3.33 36.70 0.31 0.44 nd nd nd nd 58.23
BYZ102 Yellow 0.23 0.41 6.23 38.40 0.63 0.47 0.13 nd 0.38 nd 53.11
BYZ103 Yellow 0.80 0.41 4.03 35.04 0.37 1.67 nd nd 0.44 nd 57.23
BYZ104 Yellow 0.36 0.54 4.02 36.44 0.48 1.89 nd nd 0.53 nd 55.74
 ACCEPTED MANUSCRIPT

BYZ104 Brown 0.62 1.08 6.01 34.10 0.71 1.73 0.17 nd 1.14 nd 54.44
BYZ106 Brown 0.83 1.05 4.85 33.63 0.61 1.54 0.28 nd 1.49 nd 55.71
BYZ107 Yellow 0.38 0.43 7.61 39.33 0.63 1.34 nd nd 0.46 nd 49.82
BYZ108 Yellow 0.41 0.30 6.39 38.15 0.50 1.17 0.27 nd 0.84 nd 51.97
BYZ108 Brown 0.60 0.92 7.13 33.90 0.49 1.49 0.20 nd 0.87 nd 54.39
BYZ109 Yellow 0.40 0.50 7.05 35.50 0.51 1.04 0.21 nd 0.56 nd 54.24

PT
BYZ110 Yellow 0.55 0.47 8.18 33.86 0.44 1.17 0.28 nd 0.54 nd 54.52
BYZ112 Yellow 0.41 0.84 7.55 34.40 0.55 1.77 nd nd 1.06 nd 53.44
BYZ112 Brown 0.39 1.10 8.07 34.20 0.54 1.43 nd nd 0.71 nd 53.57

RI
BYZ114 Brown 0.77 0.71 7.70 32.88 0.64 1.94 nd nd 0.58 nd 54.77
BYZ115 Cream 0.48 0.63 5.88 40.68 1.66 1.34 0.39 nd 0.58 nd 48.37
BYZ115 Brown-Red 0.45 1.84 4.20 43.42 1.73 1.01 0.15 nd 1.45 nd 45.77

SC
BYZ116 Cream 0.59 0.98 3.36 41.58 1.65 3.12 nd nd 0.61 nd 48.10
BYZ117 Cream 0.35 0.58 4.78 34.74 1.45 1.89 0.29 nd 0.68 nd 55.24
BYZ117 Brown-Red 0.36 2.64 5.56 34.70 1.50 1.76 0.41 nd 3.14 nd 49.94

NU
BYZ120 Light Green 0.46 0.52 7.81 28.87 0.77 0.33 nd nd 0.16 nd 61.08
BYZ121 Yellow 0.80 0.61 7.64 44.53 0.52 0.66 nd nd 0.36 nd 44.89
BYZ121 Brown 0.78 1.64 8.10 44.48 1.27 2.22 nd nd 1.14 nd 40.37
BYZ157 Green 0.79 0.56 3.59 38.36 1.23 1.16 nd nd 0.33 0.59 53.41
MA
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

Table A4. Qualitative PIXE composition of representative glazes

Sample Colour Elements Detected (Major and Trace) Elements Responsible for Colour
BYZ52 Dark Brown Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe
BYZ66 Blue Si, K, Ca, Fe, S, Cl, Ti, Mn, Cr, Cu Cu
BYZ69 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe

PT
BYZ70 Green Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Cu
BYZ72 Dark Brown Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu Mn
BYZ81 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe
BYZ83 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe

RI
BYZ83 Green Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Cu
BYZ87 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe

SC
BYZ89 Brown Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe (Mn)
BYZ94 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe
BYZ94 Brown Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe

NU
BYZ97 Dark Brown Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu Mn
BYZ97 White Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Ni, Cu -
BYZ98 Blue Si, K, Ca, Fe, S, Cl, Ti, Mn, Cr, Ni, Cu Cu
BYZ101 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe
MA
BYZ107 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu Fe
BYZ112 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe
BYZ117 Yellow Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Fe
BYZ120 Light Green Pb, Si, K, Ca, Fe, Cl, Ti, Mn, Cr, Ni, Cu, V Cu
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
Fig. 1

SC
NU
MA
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE

Fig. 2
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D

Fig. 3
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
Fig. 4
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE
P

Fig. 5
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D

Fig. 6
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
Fig. 7
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
Fig. 8
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE
P
CE
AC

Fig. 9
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
Fig. 10
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
Fig. 11

NU
MA
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
Fig. 12

NU
MA
D
TE
P
CE
AC
 ACCEPTED MANUSCRIPT

Table 5. Date, ware and style type of the Byzantine glazed pottery sherds.
Sample Ware type Stylistic type Date
BYZ51 Redware Chafing Dish 2nd half of 10th c. AD
BYZ52 Redware Chafing Dish 2nd half of 10th c. AD
BYZ53 Redware Chafing Dish 2nd half of 10th c. AD
BYZ54 Redware Chafing Dish / Cup (?) 2nd half of 10th c. AD

PT
BYZ55 Whiteware Red Slipped White Ware 6th c. AD
BYZ56 Whiteware Glazed White Ware II 10th – 12th c. AD
BYZ57 Whiteware Glazed White Ware II 10th – 12th c. AD
BYZ58 Whiteware Glazed White Ware II 10th – 12th c. AD

RI
BYZ59 Whiteware Glazed White Ware II 10th – 12th c. AD
BYZ60 Whiteware Glazed White Ware II 10th – 12th c. AD
10th – 12th c. AD

SC
BYZ61 Whiteware Glazed White Ware II
BYZ62 Whiteware Glazed White Ware II 10th – 12th c. AD
BYZ63 Whiteware Glazed White Ware II Late 10th – Mid 11th c. AD
BYZ64 Whiteware Glazed White Ware II (Impressed Ware) Late 10th – Mid 11th c. AD

NU
BYZ66 Whiteware Glazed White Ware ?
BYZ67 Whiteware Petal Ware 10th c. AD
BYZ68 Whiteware Petal Ware 10th c. AD
BYZ69 Whiteware Glazed White Ware II Late 10th – Mid 11th c. AD
MA
BYZ70 Whiteware Glazed White Ware II 10th – 12th c. AD
BYZ71 Whiteware Coarse Glazed Ware 10th – 12th c. AD
BYZ72 Redware Green and Brown Painted Late 11th – Early 12th c. AD
BYZ73 Redware Green and Brown Painted Late 11th – Early 12th c. AD
BYZ74 Redware Green and Brown Painted 1st half of 12th c. AD
D

BYZ75 Redware Green and Brown Painted Early 12th c. AD

BYZ76 Redware Green and Brown Painted Mid 12th c. AD
TE

BYZ77 Redware Green and Brown Painted 1st half of 12th c. AD

BYZ78 Redware Green and Brown Painted Mid 12th c. AD
BYZ79 Redware Green and Brown Painted Mid 12th c. AD
P

BYZ80 Redware Green and Brown Painted Mid 12th c. AD

BYZ81 Redware Glaze Painted Ware 13th c. AD
CE

BYZ82 Redware Glaze Painted Ware 13th c. AD

BYZ83 Redware Glaze Painted Ware 13th c. AD
BYZ84 Redware Green and Brown Painted Mid 12th c. AD
BYZ85 Redware Glaze Painted Ware (without final glazing) 3rd quarter of 13th c. AD
AC

BYZ86 Redware Dark on Light Early 12th c. AD

BYZ87 Redware Slip Painted Ware Late 11th – Early 12th c. AD
BYZ88 Redware Slip Painted Ware Mid 12th c. AD
BYZ89 Redware Slip Painted Ware Mid 12th c. AD
BYZ90 Redware Slip Painted Ware Mid 12th c. AD
BYZ91 Redware Slip Painted Ware Late 12th – Early 13th c. AD
BYZ92 Redware Slip Painted Ware Late 12th – Early 13th c. AD
BYZ93 Redware Slip Painted Ware Late 12th – Early 13th c. AD
BYZ94 Redware Dark on Light 2nd quarter of 12th c. AD
BYZ95 Redware Dark on Light 2nd quarter of 12th c. AD
BYZ96 Redware Dark on Light 2nd quarter of 12th c. AD
BYZ97 Redware Spatter Painted Ware 2nd quarter of 12th c. AD
BYZ98 Whiteware Syrian Blue Frit 12th – 13th c. AD
BYZ99 Redware Fine Sgraffito Ware (Duochrome) Early 12th c. AD
BYZ100 Redware Fine Sgraffito Ware (Duochrome) Early 12th c. AD
BYZ101 Redware Fine Sgraffito Ware Mid 12th c. AD
BYZ102 Redware Fine Sgraffito Ware Mid 12th c. AD
BYZ103 Redware Fine Sgraffito Ware Mid 12th c. AD
BYZ104 Redware Incised Sgraffito Ware 3rd quarter of 12th c. AD
BYZ105 Redware Incised Sgraffito Ware Late 12th – Early 13th c. AD
 ACCEPTED MANUSCRIPT

BYZ106 Redware Incised Sgraffito Ware Late 12th – Early 13th c. AD

BYZ107 Redware Incised Sgraffito Ware Late 12th – Early 13th c. AD
BYZ108 Redware Incised Sgraffito Ware 1st half of 13th c. AD
BYZ109 Redware Incised Sgraffito Ware 1st half of 13th c. AD
BYZ110 Redware Incised Sgraffito Ware 1st half of 13th c. AD
BYZ111 Redware Champlevé Ware 1st half of 13th c. AD
BYZ112 Redware Champlevé Ware 1st half of 13th c. AD

PT
BYZ113 Redware Incised Sgraffito Ware Late 12th – Early 13th c. AD
BYZ114 Redware Champlevé Ware 1st half of 13th c. AD
BYZ115 Redware Measles Ware 2nd quarter of 12th c. AD

RI
BYZ116 Redware Measles Ware 2nd quarter of 12th c. AD
BYZ117 Redware Measles Ware 2nd quarter of 12th c. AD
BYZ118 Redware Measles Ware 2nd quarter of 12th c. AD

SC
BYZ119 Redware Late Sgraffito Late 13th – Early 14th c. AD
BYZ120 Redware Late Sgraffito Late 13th – Early 14th c. AD
BYZ121 Redware Late Sgraffito Late 13th – Early 14th c. AD
9th -12th c. AD

NU
BYZ149 Redware Cooking Pot
BYZ150 Redware Cooking Pot 9th -12th c. AD
BYZ151 Redware Cooking Pot 9th -12th c. AD
BYZ152 Redware Cooking Pot ?
BYZ157 Redware Lamp
MA 12th c. AD
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

Table 6. Characteristics of the techniques applied for the examination of clay, glaze and clay-glaze
interface.
Technique Samples Sample Preparation Chemical Elements
analysed composition analysed
Clay FOM 71 Selected surface polished - -
SEM/EDS 71 Selected surface polished Surface Major

PT
PGAA 21 Glaze layer cut off from the Bulk Major/Trace
clay fragment
Glaze SEM/EDS 68 Cross section polished Surface Major

RI
PIXE 20 No preparation Inner matrix Major/Trace
Raman 35 No preparation - -
Interface SEM/EDS 25 Samples cut, embedded in

SC
resin and polished

NU
MA
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

Table 7. Microscopic characteristics of the different fabrics present among the samples of the
assemblage.
Ware Fabric No. of Munsell Colour Coarseness Porosity Inclusion Inclusion
samples size amount
White Coarse 2 White 10YR 8/1 Coarse Porous Small or High
medium

PT
White Fine 7 White Gley1 8/; Fine Non- Small or Average
White 10YR 8/1 porous medium
White Pink 8 Pinkish white 5YR 8/2 to 8/3; Pink Fine Non- Medium High
2.5YR 8/4 porous

RI
Red Fine 17 Reddish yellow 5YR 6/6; Fine Non- Small Low
Yellowish red 5YR 5/8 and 5/6 porous

SC
Red Coarse 20 Yellowish red 5YR 4/6 to 5/6; Fine Porous Medium High
Reddish yellow 5YR 6/6 to 7/6;
Light brown 7.5YR 5/3 and 6/3 to 6/4;
Brown 7.5YR 5/3

NU
Red Pink 14 Pinkish white 5YR 8/2; Fine Non- Medium Average
Pink 5YR 8/3; porous
Pink 2.5YR 8/4
Red Brown 7 Yellowish red 5YR 5/6; Coarse Porous Medium High
MA
coarse Light brown 7.5YR 6/4; or large
Brown 7.5 5/4
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

Table 8. Composition of major elements of the ceramic bodies estimated by SEM/EDS. The mean
group values are given as wt % (normalized to 100%); in parentheses the calculated standard
deviations.
Clay No of Na2O MgO Al2O3 SiO2 K2O CaO TiO2 FeO
Group samples
I 15 0.30 0.80 27.00 64.98 1.25 1.83 1.26 2.58

PT
(0.19) (0.36) (2.90) (2.72) (0.68) (2.12) (0.19) (0.65)
II 2 2.29 0.96 8.89 79.93 2.13 4.36 0.42 1.03
(0.47) (0.36) (0.33) (1.87) (0.10) (2.67) (0.08) (0.11)
A 17 1.77 2.68 20.13 55.24 4.26 6.84 0.86 8.23

RI
(0.39) (0.21) (0.62) (1.44) (0.37) (0.44) (0.09) (0.67)
B 24 1.04 3.45 16.84 55.66 3.50 10.29 0.83 8.39

SC
(0.43) (1.13) (1.33) (3.43) (0.60) (2.92) (0.11) (1.18)
C 14 0.88 3.07 14.39 45.98 2.88 23.97 0.82 8.02
(0.18) (0.32) (1.00) (3.47) (0.60) (4.30) (0.16) (1.09)
D 3 0.99 2.54 18.40 61.47 3.46 1.97 0.87 10.31

NU
(0.13) (1.39) (0.84) (1.12) (0.92) (0.64) (0.19) (2.51)

MA
D
P TE
CE
AC
 ACCEPTED MANUSCRIPT

Highlights
 Byzantine glazed pottery from Corinth was analysed by a multi-technique
approach.
 The manufacturing process was determined.

PT
 A sophisticated network of glazed pottery production and distribution is revealed.
 Socioeconomic changes (early 13th c. AD) had significant impact on the glazed

RI
pottery production.

SC
NU
MA
D
P TE
CE
AC