Sancta Sophia, Trebizond: A Note on the Materials and Technique

Author(s): Joyce Plesters

Source: Studies in Conservation , Nov., 1963, Vol. 8, No. 4 (Nov., 1963), pp. 131-135
Published by: Taylor & Francis, Ltd. on behalf of the International Institute for
Conservation of Historic and Artistic Works

Stable URL: https://www.jstor.org/stable/1505229

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/1505229?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Taylor & Francis, Ltd. and International Institute for Conservation of Historic and Artistic
Works are collaborating with JSTOR to digitize, preserve and extend access to Studies in
Conservation

This content downloaded from

139.179.182.150 on Sun, 10 Jan 2021 07:09:40 UTC
All use subject to https://about.jstor.org/terms
 JOYCE PLESTERS

Sancta Sophia, Trebizond-

A Note on the Materials and Technique
Received 14/6/63

A NUMBER Of samples of Turkish plaster and consists of fine white calcium carbonate with
whitewash to which fragments of the original only occasional grains of brownish sand.
Byzantine wall painting adhered were exam- Layer (b) is the thick Turkish plaster applied in
ined in the Scientific Department of the the nineteenth century. It is mainly calcium
National Gallery, London. During the course carbonate, coarse and of a cream colour, but
of the work at Trebizond Mr Winfield collec- speckled throughout its thickness with small,
ted nearly one hundred small samples of paint hard, glassy, dark particles which proved to be
and plaster, and some of the more recently the local sand from the beach at Trebizond.
collected still await detailed examination, Examined under the microscope at low power,
so that the following is only a preliminary these sand grains, blackish-green by reflected
report. It is hoped that a fuller account may light, are a clear olive green by transmitted
appear elsewhere at a later date [I]. light. Chemical analysis showed their com-
position to be an iron magnesium silicate. The
The Covering Layers dark green mineral which forms the main con-
stituent of this sandwas identified as a diopsidic
Fig. I represents the layer structure of paint pyroxene, the only other constituent present,
and plaster before the uncovering of the paint- and this in minor quantities only, being a basic
ings. Of the covering layers, layer (a) is white- plagioclase [2]. The identity of the dark green
wash with which the paintings were coated sandy particles is of interest because, as will be
--(c) 19th-century seen below, in two samples of the original
whitewash
Byzantine paintings it was found to have been
* -(b) 19th-century
AS 4 ~a Turkish plaster Coat- used as a pigment in underpainting. A con-
4 * *. (a) i6th-century ing siderable amount of fibrous material was also
/ whitewash Layers
-(3) Paint layers - found in the Turkish plaster (layer (b)). Two
\(2) Upper Byz- types of fibrous material were present, the first
antine plaster Orig-
fine and rather silky and identified as flax fibre,

V " --(I) Lower Byz- Layersthe

antine plaster _j
other coarser and more brittle and looking
like fragments of grass stems, although the
exact botanical origin of this coarser fibre
FIG I. The layer structure of paint and plaster at
Trebizond before the paintings were uncovered. could not be established [3]. In many of the
samples this Turkish plaster is much harder and
when the church was transformed into a more compact than the original Byzantine
mosque in the sixteenth century. It is not plaster, and its adhesion and the adhesion of
present in all areas, and parts may have fallen the sixteenth-century whitewash to the top
away between the sixteenth and the nineteenth surface of the Byzantine paint seems better
centuries (when the walls were replastered). It than that of the paint to the Byzantine plaster
131

This content downloaded from

139.179.182.150 on Sun, 10 Jan 2021 07:09:40 UTC
All use subject to https://about.jstor.org/terms
 132 Joyce Plesters
below. But, as mentioned
pared from below,
many of the samplesthisin orderconditio
to
examine theexamined
is peculiar to the samples layer structure under the micro-
rather tha
scope. The pigments were
characteristic of the paintings asidentified by both
a whole. Laye
optical
(c) is the topmost layer and chemical tests carried outwith
ofwhitewash under the which
the nineteenth-century microscope on both the cross-sections
plaster and the
was finished.
unmounted
consists of a thin layer of fragments.
fine white calcium
Solubility, staining,
carbonate with practically no and inclusions.
combustion tests A
three layers contain awere carried out of
trace on scrapings of the various ion, bu
chloride
this could have originated
paint layers in order from
to establish if anythe
organic use o
binding
unwashed sea sand in themediumTurkish
was present. Most ofplaster
the o
layer (b). samples gave a negative result, but from some
of the thicker, uppermost paint layers a little
glue-like material could be extracted by re-
The Byzantine Plaster
peated treatment with hot water. Chromato-
In most of the samples examined the original graphic analysis carried out on these samples
Byzantine plaster was missing, the paint confirmed the presence of a small proportion
having come away with the Turkish plaster of animal glue [4]. Unfortunately it is not pos-
and whitewash. The reason for this is that in- sible to say with certainty whether this is the
evitably in a few places paint came away with remains of the original painting medium or
the Turkish plaster and whitewash and it was the remains of glue applied in some later
thought better to use these pieces for examin- treatment of the frescoes. MrWinfield's study
ation where possible rather than to cut fresh of the techniques of the paintings themselves
samples. suggests that the broad underpainting of the
Of the original layers shown in Fig. I, figures, generally carried out in a yellow ochre
layer I is the undercoat of Byzantine plaster. It colour, might have been done in a fresco tech-
isofacreamy-buffcolour (ferric iron is present) nique, whereas the upper paint layers, such as
and contains a great many fragments of coarse the modelling of the draperies and final details,
straw as a binder. Very few sand particles were could have been done in a glue distemper. The
present, but these included a few grains of the absence of glue medium in some of the paint
local blackish-green sand, and there were also samples does not, of course, preclude its pre-
a few particles of brick dust. Layer (2) is the sence in the paint initially, for it could have
final layer of Byzantine plaster, a thin coat of been leached out by water or destroyed by
pure white calcium carbonate with virtually moulds or bacteria over the years, especially
no solid inclusions. Both these layers also con- in the humid conditions which prevailed in
tain a trace of chloride. the church at Trebizond.

The Paint Layers and the Medium The Pigments

In many parts of the Byzantine painting
there are two or even three superimposed
paint layers, and these are by no means mere
overlappings of different-coloured areas of the
design but intentional shading and modelling
of areas of colour, and in some cases superim-
posed layers are used for the purpose of pro-
ducing colours unobtainable with the single
pigments available. Cross-sections were pre-
The most frequently occurring pigments
were, as might be anticipated, the earth
colours. As mentioned above, a considerable
amount of underpainting is carried out in a
yellow ochre colour. A very intensely col-
oured red-brown earth also occurs, corres-
ponding roughly to the colour we know as
burnt sienna but stronger and redder in hue
than most specimens which go under that

This content downloaded from

139.179.182.150 on Sun, 10 Jan 2021 07:09:40 UTC
All use subject to https://about.jstor.org/terms
 Sancta Sophia, Trebizond-A Note on the Materials and Technique I33
name today. A dark red ferric oxide (haematite)
is used for purplish and crimson hues, though
some of these areas may now appear more
purplish, or rather dullish mauve, than origin-
ally intended owing to the efflorescence of
crystalline calcium carbonate on the surface of
the paint (in other areas where purple was re-
quired, superimposed layers of the red-brown
earth and ultramarine are used).
Green earth (or terre verte, a hydrosilicate of
iron, magnesium, aluminium, and calcium) was
the only green pigment identified in the whole
range of samples. Mr Winfield has described
how this green pigment has been used for the
underpainting of the flesh in the traditional
manner characteristic both of Byzantine paint-
ing and that of the Italian tre- and early
quattro-cento. This pigment is also used for
trees and drapery. Often green earth is of a
rather greyish-green hue, but some of the
samples were of such a strong slightly bluish
green colour as to be mistaken at first sight for
malachite. Elsewhere the colour of the green
earth has been modified, quite deliberately it
seems, by the use of different-coloured under-
paint. For example, a colour close to leaf green
has been achieved by using a layer of green
earth over one of yellow ochre, and a very dark
green by underpainting with carbon black.
One pigment present which is rarely used in
fresco painting is vermilion (mercuric sul-
phide), used, for example, to produce a bril-
liant range of reds and pinks of an angel's wing
feathers, well-preserved and showing no signs
of the blackening which sometimes affects this
colour when used on wall-paintings. A little
vermilion is also used in the final touches of
the flesh tones, although in general the warm
tones of the flesh are in earth colours.
The only white pigment identified in the
whole range of samples was lime white (now
present as calcium carbonate), and this is used
extensively for highlights on flesh and drapery
and, as Mr Winfield has described, for white
inscriptions. In some cases it is applied quite
thickly, as in the case of pearls on the costume
of an angel.
This content downloaded from
139.179.182.150 on Sun, 10 Jan 2021 07:09:40 UTC
All use subject to https://about.jstor.org/terms
The black pigment is a form of carbon black
which, from the splinter-like fragments which
it contains, is presumably a charcoal black.
This black is sometimes used on white to give
the impression of a blue colour, and some
samples labelled 'blue paint' contained only
carbon black and white.
Rather surprising was the extensive use of
natural ultramarine, which was the principal
blue pigment throughout the series, and
used even for the considerable areas of blue
background. As in the case of green earth, this
single pigment was used in different ways to
obtain different colour effects. In some of the
paler blue backgrounds a 'dilute' form of the
pigment is used, having only small blue par-
ticles and a high proportion of colourless
material including calcite. This roughly corres-
ponds to what we know as ultramarine ash, i.e.
the last and palest fraction obtained in the
traditional method of extraction of ultra-
marine from lapis lazuli.Where a more bril-
liant blue was required the ultramarine has
been coarsely ground and thickly applied, but
still contains a high proportion of colourless
material. This might suggest that the pigment
was not extracted by the method of mixing
the ground mineral into a molten mass of
waxes and resins, then kneading the mass
under aqueous alkaline solution (a method first
practised in Italy by about I200 and which
continued to be used unchanged up to modern
times), but by some cruder method. The very
dark blue of some of the backgrounds to the
figures is produced by carbon black underpaint
with a thin top layer of natural ultramarine.
Samples from one area only had gilding, and
these were from the conch of the apse. The
gilding was done with gold leaf and must have
been carried out with a water, not oil, mor-
dant. Where used for the lettering ofthe Mother
of God the gold leaf was laid directly on the
white plaster, and where used for her halo it
was laid on a layer of yellow earth colour.
From the conch of the apse also came the
only sample in which was identified a blue
other than ultramarine, namely azurite (a basic
134 Joyce Plesters
copper carbonate mineral).
glue A small
could bearea of it
extracted, al
occurred adjacent to and partly
tain thatunder
this is the
the blue
origin
lettering. Perhaps the oddest featureto
With regard of the
the pigm
samples examined was the usethe
series in two
earth instances
colours pl
of a thin layer of the fine
The blackish-green
green pigment local
most
sand (described under The
bothCovering
series is Layers) as
green earth
.amii the
the pigment of underpaint, series a brighter
topmost layer gr
in both cases being of green
since earth.
this was It is
in excep-
areas wh
tional for a painter to blue
depart would
fromhave been ex
the con-
cludedbut
ventional range of pigments, that the green
certainly thecolo
Trebizond sand is an attractive and unusual
the conversion of the blue
a green
material, although it must have chloride
proved grittycompou
and unsympathetic to thepresent
brush andin the plaster.
its use as a Az
pigment was obviously only
not persevered
one of the with.Trebizond
One observation which conchmight
of the be made
apse, and its
changed, despite
about the method of employment of the pig- the pr
ments mentioned above is that of
amount nowhere
chloride isin
a the
physical mixture of two milion
or moreoccurs in both
coloured pig-ser
ments used, although someblack pigment
coloured is a carbon
pigments
are present mixed with to be white.
lime charcoal
Theand the onl
indivi-
dual pigments are either lime white.
used pure or Gilding (with
the desired
colour effect obtained by superimposed
only in the conch layers.
of the a
and the
It is interesting to compare thematerials
only sample
and fro
techniques of the Trebizond
includedfrescoes with
gold, also as gold
those of another seriesconch
of lateof Byzantine wall
the apse, though f
were also and
paintings of which an admirable present on hal
detailed
account was given in thisAt journal
Trebizond a fewtheyears
predom
was natural
ago [5]. These are the frescoes in the ultramarine
church
now known as the Kariye qamii ?amii
it was atazurite.
Istanbul,
The most unexpected
and date from the early fourteenth century. f
Like those at TrebizondS amii
the was the presence
Kariye ;amii o
frescoes were hidden by various
ment made coatings
from a after
cobalt
the church became a mosque, but have
in the earliest partnowof the
been uncovered and restored
dating [6].
from Two between
distinct ca. 1
probably
layers of Byzantine plaster havethebeen earliest
used in recor
each series, and a photograph
pigment of the under
in European pain
found 9amii
layer of plaster from Kariye in the Trebizond
shows it to sam
be quite similar to that any indicationboth
at Trebizond, of the pres
con-
taining a great deal of is
coarse straw,
reported asbut con-
occurring
taining no sand at all conjunction
at Kariye Camii with and a
the smal
negligible amount in theResults
samples of the detailed
from Trebi- ex
zond, and no marble dusttwo
in either.
series of late Byzantin
provide
Turning to the paint layers, a useful
there link wi
was only
slight evidence of the presence
quattrocentoof an organic
fresco painti
binding material in the samples from Kariye
JOYCE PLEST
9amii, but from one or two of
National the thickest
Gallery
paint layers of the Trebizond samples
London, IW.c .2. a little

This content downloaded from

139.179.182.150 on Sun, 10 Jan 2021 07:09:40 UTC
All use subject to https://about.jstor.org/terms
 Sancta Sophia, Trebizond---A Note on the Materials and Technique 135

Notes and References

I In a forthcoming publication on S. Sophia, Trebizond, by Professor D. Talbot-Rice.

2 I am indebted to Dr R. W. Le Maitre of the Department of Mineralogy, British
Museum (Natural History) for the identification of the mineral forms of the
components of the sand.
3 I am indebted to the Department of Biomolecular Research, University of Leeds,
for the examination of the fibrous plant material from the Turkish plaster.
4 Paper chromotography carried out by Miss M. Hey of the National Gallery Lab-
oratory.
5 RUTHERFORD J. GETTENS and GEORGE L. STOUT, ' A Monument of Byzantine Wall
Painting-the Method ofConstruction ', Studies in Conservation, 3, no. 3 (I 958),
pp. 107-9.
6 CARROL WALES, 'The Treatment of Wall Paintings at the Kariye 9amii ', Studies
in Conservation, 3, no. 3 (1958), pp. 120-3.

SANCTA SOPHIA, TREBIZONDE

NOTE SUR MATERIAUX ET METHODES

Resumre'

Ceci est un rapport prdliminaire sur fut le seul pigment vert identifid), blanc de
l'examen microscopique et chimique d'dchan- chaux, noir de charbon de bois, vermillon,
tillons de peinture et de platre. Au XVIe outremer naturel (qui, a une petite exception
siecle, les Turcs recouvrirent les fresques prIs, est le seul pigment bleu identifie). On
d'une couche de chaux, suivie au XIXe trouve des dorures seulement dans les pr6-
siecle d'une 6paisse couche de platre et livements effectu s dans la vofite d'abside.
encore de la chaux. La chaux employee Un trait inusit6 est l'emploi de sable local vert-
consiste en carbonate de calcium relativement noiratre comme premi&re couche pour du
pur. Mais l'6pais plktre turc contient, en plus, vert pale. Les pigments n'Itaient pas m lang6s
du sable local vert-noir$tre, du lin et autres exceptd avec du blanc, mais des effets de
fibres v6g6tales. Le platre byzantin, sous la couleurs varids 6taient obtenus par des
peinture, est 6galement du carbonate de couches superpos6es.
calcium, contenant de la paille, mais pas de Cet article pr6sente une comparaison avec
sable. un autre groupe de peintures murales byzan-
Dans les couches de revetement inf6rieures, tines plus r6centes, celles de la Kariye Qamii,
on ne put d6couvrir aucun support organique, " Istanbul, d6crites prec6demment dans cette
mais un peu de glu animale fut extraite des revue.

6paisses couches superieures. Les pigments J. P.

presents comprennent des terres (la terre verte

This content downloaded from

139.179.182.150 on Sun, 10 Jan 2021 07:09:40 UTC
All use subject to https://about.jstor.org/terms