(BEING CONTINUED FROM 28/08/11)
Dyestuffs and Biological Sources
Natural dyestuffs identified in the extracted samples are summarized in table 1. The results
are presented with respect to the macroscopic colours of the samples and are based on the
detection of specific colouring compounds which were used as indices for the identification of
the dyestuff sources. Specifically, carminic acid was used as an index for the identification of
cochineal, a brazilein derivative (Bra´) and type C compound were indicative for the use of a
soluble redwood (Nowik, 2001), laccaic acid A and indigotin were used as indices for the
identifications of lac (Kerria lacca Kerr) and an indigoind dye source, respectively. Madder
components such as alizarin and purpurin were detected in samples I.2 and III.2. Sulfuretin
was used for the identification of young fustic (Cotinus coggygria Scop.).
Finally, the identification of dyer’s broom (Genista tinctoria L.) was based on the detection of genistein,
apigenin and luteolin. It is noteworthy, that luteolin was detected not only in samples in which
dyer’s broom was found (table 1) but also in samples III.1 and III.2 along with several other
colouring compounds, none of which, however, could be used for the assessment a particular
dye source which could explain the presence of luteolin. The red colour of these two samples
suggests that a luteolin-based yellow dyestuff was used during dyeing in limited quantities.
This might be the reason for not detecting any secondary colouring compound which could be
used for the identification of a luteolin-based dyestuff.
The chromatogram collected for sample III.1 is provided in figure 3 as an example. In the
following the discussion focuses on dyestuffs for which the assessment of a specific
biological source cannot be achieved directly from the analytical results. These dyestuffs are
the indigoid dye source, the soluble redwood, cochineal and madder. Lac, young fustic and
dyer’s broom clearly correspond to Kerria lacca Kerr, Cotinus coggygria Scop. and Genista
tinctoria L., respectively.
It is known that indigotin is contained in woad, Isatis tinctoria L. and in indigo, (Indigofera
species e.g. Indigofera tinctoria L.). Woad and indigo were known since antiquity and have
similar compositions. It is not possible to separate woad from indigo using historical or even
chemical-analytical information, despite the use of highly sensitive techniques such as the
HPLC. Consequently, in table 1 we only report the presence of an indigoid dye source which
could be either indigo or woad. Any further comment on this rather general result is not
The redwood dyestuff reported in table 1 belongs to the so called “soluble redwoods”. The
latter correspond to the family of the Caesalpinia trees, which includes Caesalpinia sappan
(native to Asia), Caesalpinia etchinata (native to America) and several others. The various
species of redwoods (many times are called brazilwoods) cannot be easily distinguished on
the basis of their chemical compositions. To some extent this was achieved by Nowik (2001).
It is noteworthy though that this study focused exclusively on reference redwood samples i.e.
plant extracts of known origin. In our case, unknown samples extracted from historical
artworks were investigated; the assessment/identification of the exact redwood plant was not
Cochineal is obtained from three scale insects: (i) Mexican cochineal is obtained from
Dactylopius coccus Costa, (ii) Armenian cochineal from Porphyrophora hameli Brandt and
(iii) Polish cochineal from Porphyrophora polonica L. Mexican cochineal is native to
America and apparently it was used after the discovery of the new world. Consequently, its
use in artworks created before the 15th century should be excluded. In the present study
cochineal was identified in objects IV (samples IV.3 and IV.4) and VI (sample VI.1). Both
objects are considered to be of the 18th century and therefore the use of Mexican cochineal
during dyeing is possible. Armenian and Polish cochineal were known since ancient times. In
principle, the identification of the specific source of cochineal (Dactylopius coccus Costa,
Porphyrophora hameli Brandt, Porphyrophora polonica L.) can be performed on the basis of
the relative composition of the minor constituents of cochineal, accompanied by a detailed
statistical analysis (Wouters and Verhecken, 1989).
The results obtained from the analyses of reference (known origin) cochineal samples are quite clear and the distinction of the three species of cochineal can be obtained. The application of this analysis in historical samples,
however, does not always lead to robust conclusions with respect to the exact source of
cochineal. In historical samples several difficulties such as (i) the usually small available
amounts of samples and (ii) the degradation of dyestuffs developed because of ageing effects
and/or extensive use of the historical object, raise sometimes a considerable degree of
uncertainty regarding the cochineal source found in the samples. We report that carminic acid
was detected in sample IV.3 in small quantity. It was therefore impossible to detect any minor
cochineal component. Consequently, sample IV.3 is not included in the following discussion;
we focus only on the results obtained for samples IV.4 and VI.1. Distinguished (high) peaks
which corresponded to carminic acid were clearly recorded in the chromatograms of both
IV.4 and VI.1 samples.
The relative composition of four components is used as a rule for the identification of a
cochineal source: dcII, carminic acid (CA), kermesic acid (KA) and flavokermesic acid (FL).
According to Wouters and Verhecken (1989) Porphyrophora polonica L. (Polish cochineal)
contains KA and FL at elevated concentrations. Performing measurements at 275nm the peak
area of KA+FL should be within 12-38%. The remaining 62-88% corresponds to CA; dcII is
contained in trace (Wouters and Verhecken). KA and FL were not detected in sample VI.1.
This can lead to the speculation that the use of Polish cochineal during dyeing can be
excluded. A similar conclusion can be drawn for sample IV.4.
In this case KA+FL was found on the order of 2%, which is substantially lower than the corresponding composition (12-
38%) reported for Polish cochineal samples (Wouters and Verhecken). We focus now on the
two remaining candidates: Dactylopius coccus Costa (Mexican cochineal) and Porphyrophora hameli Brandt (Armenian cochineal). The distinction of the two can be done on the basis of the relative composition of dcII and CA. Mexican cochineal contains dcII (1.4-3.8%) at elevated concentrations in contrast to Armenian cochineal in which the concentration
of dcII ranges from 0.1 to 1.2% (Wouters and Verhecken). The dcII compound was not
detected in any of the samples IV.4 and VI.1. Therefore this is a strong indication that the use
of Armenian cochineal during dyeing is more likely than the use of Mexican cochineal. In
summary, we note the HPLC results provided indications that Armenian cochineal (Porphyrophora hameli Brandt) was most likely the dyestuff source used for the treatment of samples IV.4 and VI.1. However, this should not be considered as a robust result (only as a speculation) because degradation mechanisms developed in the historical samples might have
altered the composition of the cochineal originally used by the dyers of the garments.
There are several madder species around the world. In the Mediterranean area, however, two
species are dominant: Rubia tinctorum L. and Rubia peregrina L. Our discussion will therefore start with the assumption that only these two species should be considered as the dyestuff sources detected in samples I.2 and III.2. Previous studies, performed by Wouters (2001) suggested that detailed measurements of the relative areas of the peaks, which
correspond to the colouring components of madder, are necessary for the identification of the
exact madder source.
Such quantitative measurements, performed at 255nm, have shown that the relative peak area ratio of alizarin versus purpurin is around 1.19 for samples extracted from yarns dyed with Rubia tinctorum and around 0.16 for samples obtained from yarns dyed
with Rubia peregrina. For the samples of interest (samples I.2. and III.2) we report that: (i)
alizarin was found in both samples, along of course with purpurin. (ii) The relative peak area
ratios of alizarin versus purpurin for samples I.2 and III.2 were found to be 1.67 and 0.84,
respectively. These measurements were performed at 250nm which is relatively close to the 255nm used by Wouters. The detection of alizarin at elevated concentrations may lead to the conclusion that the madder dye found in samples I.2 and III.2 was obtained most likely from Rubia tinctorum L.
Colours and Hues
In this paragraph we try to correlate the analytical results with the macroscopic colours of the
samples. It is important to note that the colours of the samples reported in table 1 do not
coincide necessarily with the colours that these textile samples used to have when they were
dyed. Colour fading occurring because of ageing effects and/or extensive use of the object
might have altered the original colours of the fibres. On the other hand some interesting
comments can be done on the results of table 1, as follows.
The red colour was obtained from several combinations. For example the combination of
redwood and the blue indigoid dyestuff resulted in the production of dark red (sample III.1 of
figure 2). Madder was used to produce a lighter red colour (e.g. sample III.2 of figure 2).
Other reddish samples contained redwood (sample VII.1), mixture of cochineal and redwood
(sample IV.4) and mixture of lac and madder (sample I.2). Three samples which appeared to
be yellow contained the same dyestuffs: dyer’s broom in mixture with young fustic. This result corresponds to samples III.5, IV.1 and IV.2. Young fustic was found to be the exclusive colouring matter in samples II.1 and V.1. Finally, samples IV.3 and VI.2 appeared to be macroscopically yellow.
However, traces of red (cochineal, redwood) and blue (indigoid dye) were detected in these samples. The amounts of the red and blue components were small according to relative ratios recorded at the HPLC chromatograms. Therefore their presence
does not affect substantially the macroscopic (yellow) colour of these samples. Two samples
(III.3 and III.4) appear to be green. Both were dyed with the yellow dyer’s broom, and a blue
indigoid dye. The orange colour of sample VI.1 was achieved by the combination of reddish
(cochineal and redwood) dyes with a yellow one (dyer’s broom). Finally, sample I.1 which
appeared to be brown contained dyestuffs of all basic colours i.e. red, yellow and blue.
COMPARISON OF DYESTUFFS USED IN GARMENTS AND ICONS
Table 2 summarizes the dyestuffs detected in the samples which were extracted from the
seven historical garments. For comparison, dyestuffs found in samples extracted from fifteen
icons are included (Karapanagiotis et al., 2007). We note that the two artwork collections
(garments and icons) have the same provenance (area of Chalkidiki, Greece), they correspond
to the same historical period (icons: 14th – 19th cent.; & garments: late 15th – 19th cent.) and
both are related to the Byzantine and post-Byzantine tradition and art. Table 2 suggests that
similar red and blue dyestuffs were used by the dyers and iconographers of the artworks.
However, yellow organic dyes were not detected in any of the tested icons, in contrast to the
garments in which young fustic and dyer’s broom were identified. Although, yellow dyes
were known by the dyers and therefore iconographers, the latter preferred to use gold sheets
and inorganic pigments (e.g. yellow ochre) to produce the yellow colour.
The following organic dyes were identified in samples extracted from seven historical
garments (15th – 19th cent.) of Mount Athos (Chalkidiki, Greece): dyer’s broom (Genista
tinctoria L.), young fustic (Cotinus coggygria Scop.), an indigoid dye source either indigo
(Indigofera species) or woad (Isatis tinctoria L.), madder (most likely Rubia tinctorum L.),
cochineal (probably Porphyrophora hameli Brandt) and lac (Kerria lacca Kerr). Also, the
identification of a brazilein derivative indicates the presence of a Caesalpinia dye source in
the samples. Similar red (except lac) and blue dyes were identified in samples extracted from
icons of the same historical period (and same provenance). On the contrary yellow dyestuffs,
commonly found in garments, were not detected in any of the icon extracts.
Support by the European Commission through the European project INCO CT 2005 015406
MED-COLOUR-TECH (www.medcolourtech.org) is gratefully acknowledged. C. Karydis
would like to thank the Holy Synod of the Greek Orthodox Church for funding his PhD.
Ioannis Karapanagiotis1, Christos Karydis2
1ORMYLIA Art Diagnosis Centre, Chalkidiki 63071, Greece
2University of Lincoln, Conservation Department, Chad Varah House,
Wordsworth St. Lincoln. LNI 3BP, UK
1. Bal, M., and Bryson, N. 1998. Semiotics and Art History: A Discussion of Context and
Senders. In: D. Preziosi, ed. The Art of Art History: A Critical Anthology. Oxford: Oxford
University Press, 242-262.
2. Braun, J. 1907. Die Liturgisle Gewandung in Occident und Orient. Freiburg: St Louis Mo.
3. Chatzidaki, E. 1953. Ecclesiastical Embroideries. Athens: Benaki Museum.
4. Dix, D.1945. The Shape of the Liturgy. Westminster: Dacre Press.
5. Johstone, P. 2002. High Fashion in the Church. Leeds: Maney.
6. Karapanagiotis, I., Karydis, C., Laka, A and Panagiotou, C. 2006. Identification of Dyes
on Ecclesiastical Garments from the Holy Mountain of Athos: The Sakkos of Emperor
Ioannis Tsimiskis. 25th Meeting of Dyes in History & Archaeology, 20-23 September,
7. Karapanagiotis, I., Valianou, L., Sist. Daniilia and Chryssoulakis, Y. 2007. Organic Dyes
in Byzantine and Post-Byzantine Icons from Chalkidiki (Greece). Journal of Cultural
Heritage, 8 (3): 294.
8. Karydis, C. 2006. Introduction to the Preventive Conservation of Textiles. Athens: Futura.
9. Kourkoulas, K. 1960.The Priestly Garments and their Symbolism in the Orthodox Greek
10. Nowik, W. 2001. The possibility of differentiation and identification of red and blue
“soluble” dyewoods. Dyes History Archaeology, 16/17:129.
11. Wouters, J. 2001. The dye of Rubia peregrina – I. Preliminary investigations. Dyes
History and Archaeol, 16/17:145.
12. Wouters, J., Verhecken, A. 1989. The coccid insect dyes: HPLC and computerized diodearray
analysis of dyed yarns. Studies in Conservation, 34:189.
SOURCE 9th International Conference on NDT of Art, Jerusalem Israel, 25-30 May 2008