The terracotta army from the burial complex of the first Chinese
Emperor Qin Shihuang (died 210 BC) is well known as a collection of
grey sculptures which appear from the distance to be a uniform troop. All
the sculptures, chariots and weapons were originally painted, but the few
colours which survived the centuries were lost during excavation. One of
the aims of the Chinese-German co-operative project is the difficult task
of conserving this polychromy. Parallel to conservation efforts, analyses
of the pigments and their application techniques, and reconstructions of
the colour combinations and the elaborately painted patterns, have been
undertaken. They have helped provide an understanding of the poly-
chromy as it was originally intended. The polychrome schemes are of
extraordinary quality and importance. They provide insight not only into
ancient painting techniques, but also into the style of clothes and uniforms
in the ancient state of Qin. Reconstructions of the polychromy on two
replicas are an attempt to visualize the original appearance of the
Die Terrakottaarmee aus der Grabanlage des Ersten Chinesischen Kaisers
(gest. 210 v. Chr.) ist gut bekannt als Formation grauer Figuren, die aus
der Ferne als einheitliche Truppe erscheinen. Ursprünglich waren jedoch
alle Skulpturen und auch die Wagen und Waffen bemalt. Die Farbfas-
sungen haben leicht reduziert die Jahrhunderte überstanden, gehen aber
bei der Ausgrabung unwiderruflich verloren. Einer der Schwerpunkte des
Chinesisch-Deutschen Forschungsprojektes ist die schwierige Aufgabe,
diese Farbfassungen zu erhalten. Parallel dazu werden Untersuchungen
zu Pigmenten und Maltechnik, zur Rekonstruktion der Farbschemata
und der aufwändig gestalteten Muster wurden durchgeführt, die hier
vorgestellt werden. Die Farbfassungen sind von außerordentlicher
Qualität und großer Bedeutung. Sie erlauben einen Einblick nicht nur
in die antike Maltechnik, sondern auch in den Stil der Kleidungen und
Uniformen des antiken Staates Qin. Rekonstruktionen der Fassungen auf
zwei Gipskopien sollen versuchen, sich dem ursprünglichen Aussehen
der Figuren zu nähern.
The terracotta ar my is one of the most important archaeological
discoveries of the twentieth century and one of the most famous
cultural monuments and greatest tourist attractions in China.
The terracotta army is part of the burial gifts of the first Chinese
Emperor Qin Shihuang. Twelve years after unifying the country,
Qin Shihuang was buried close to his capital city in 210 BC. It
has always been known that his grave is under a 70 m high earth
pyramid near Lintong, 30 kilometres east of Xi’an, the capital
of Shaanxi Province, but the extent of the complex and the qual-
ity and quantity of the treasures concealed underground were
unknown. In 1974, peasants discovered the terracotta army by
accident. This resulted in an excavation campaign which has
lasted for more than 30 years: 188 pits surrounding the grave
mound are known. They contain objects representing all aspects
of the emperor’s life: chariots, civil servants, acrobats, an under-
ground river with bronze birds, stables, a zoo etc. Four of these
pits are occupied by the terracotta army.
This gigantic burial complex eclipsed all previous royal graves
and set standards for the following emperors. From this time
onwards, a terracotta army was among the burial objects in every
imperial grave. Qin Shihuang’s terracotta army however was not
only the first of its kind but also the only one with figures that
are larger than life-size.
But Qin Shihuang overextended himself with his huge build-
ing projects. Moreover what was planned as an eternal resting
place did not remain undisturbed for long. According to the
annals written in the Han Dynasty around 90 BC [1–3], the
burial complex was never completed and was destroyed by a
A COLOURFUL WORLD FOR THE EMPEROR’S SOUL: THE POLYCHROMY OF
THE TERRACOTTA SCULPTURES AT QIN SHIHUANG’S BURIAL COMPLEX
Catharina Blänsdorf and Xia Yin
rebel army in 206 BC. Archaeological research seems to confirm
this, as several of the pits are empty and all show damage from
a severe fire.
The initial concept of creating an underground world with
thousands of larger than life-sized terracotta figures is not well
understood, though burial figures have been found in tombs of
the sixth century BC. Ideas on the intent behind the army have
mostly focused on the clay as an important material: ter racotta
makes mass production possible, but also allows detailed mod-
elling and portrait-like faces. The terracotta army shows this
fascinating combination of mass and individuality. This impres-
sion is intensified by its present appearance: rows of greyish
brown figures seem to rise out of the ground between ear then
walls of the same colour. Individual faces and types can be per-
ceived but the monochromatic coloration dominates, emphasizing
the impression of unity and uniformity. The grey seems appropri-
ate for an underground ‘Army of the Netherworld’.
But the terracotta figures were originally completely painted.
This has always been known, since extensive remnants of poly-
chromy were found. This was not surprising because realisti-
cally painted grave figures of servants and guards, equipped with
clothes and weapons, exist from the fourth century BC [4, 5].
When excavations in Lintong began, it was also known that later
terracotta armies from the Han Dynasty (206 BC–AD 220) were
colourfully painted, as for example the 2400 figures from the
burial complex of the first Han Emperor (died in 195 BC) in
Yangjiawan, excavated in 1965. But the polychromy of Qin
Shihuang’s terracotta figures flakes off directly after the
surfaces of the figures are exposed to the damp loess soil. Only
tiny remnants of pigment remain on the terracotta. So it is
difficult to imagine a coloured army instead of a grey one.
THE CHINESE-GERMAN RESEARCH PROJECT
The Chinese-German Research Project for Preservation of the
Cultural Property of Shaanxi Province was founded in 1988.
It is supported financially by the German Federal Ministry for
Research and Education (BMBF) and the Chinese State Admin-
istration of Cultural Heritage (Zhongjia wenwu ju). Two German
and three Chinese institutions are involved: the Bavarian State
Department for Historical Monuments (BSDHM, Bayerisches
Landesamt für Denkmalpflege) in Munich; the Roman-Germanic
Central Museum (Römisch-Germanisches Zentralmuseum)
in Mainz; the Museum of the Terracotta Warriors and Horses
(MTW) in Lintong; the Archaeological Institute of Shaanxi
Province in Xi’an; and the Technical Center for the Conservation
and Restoration of the Cultural Heritage of Shaanxi Province
The aim of the project is to find solutions for unusually dif-
ficult conservation problems, accompanied by research on objects
which all are of great cultural importance and high artistic qual-
ity. Natural sciences and academic research are combined with
conservation in the project: restorers, chemists, geologists, art
historians and sinologists work together. Furthermore, contacts
and co-operation with other institutions such as universities,
research institutions and chemical companies give wider possi-
bilities for scientific exchange. The BSDHM and the MTW have
focused together on problems in the preservation of the terracotta
army and newer findings from Qin Shihuang’s burial complex:
conservation of ear then structures, assembling the broken terra-
cotta figures, climate control and microbiological problems and,
since 1999, conservation of the newly-discovered stone armour
sets made of limestone plates.
Conservation of the polychromy of the terracotta sculptures
This has been an important objective of the project from the start.
For a long time no possibility of conserving the polychromy was
seen in Lintong: the paint layers on the sculptures were lost dur-
ing excavation. Some could be rescued in the soil, as the pigment
layer adhered to the soil on top of the sculptures — as a kind
of ‘negative imprint’. Because colour photography was limited,
colours were mainly recorded briefly in lists, and individual
patterns were documented in drawings.
The critical and most problematic step for the preservation of
the polychromy is the conservation of the brown ground layer. It
was obviously the material which on drying developed cracks,
deformed and peeled off the terracotta, resulting in the complete
loss of ground and pigment. One of the first results of the research
project was the identification of the material of the ground layer:
Though severely degraded, it could be identified as east Asian
lacquer (qi in Chinese, urushi in Japanese). The conservation
of the water-saturated lacquer is a problem which required the
development of a new conservation technique, as all standard
conservation methods had failed. Principally, a non-evaporating
material must be introduced to replace the water and a consolid-
ant is needed to re-establish the adhesion between lacquer and
terracotta. The extremely fine porosity greatly limits the pos-
sibilities and excludes the adhesives which are commonly used
in conservation. Two procedures suitable in principle have been
developed: polyethylene glycol (PEG) 200 and a water-soluble
dispersion of consolidant, or methylacrylate or methacrylate
monomers (HEMA) polymerized by electron beam irradiation,
published elsewhere [6−11]. Each involves disadvantages and dif-
ficulties; thus improvements in these procedures and long-term
tests are still in progress. The focus of the present text is on the
results of current research on painting techniques.
The polychromy on newly-excavated sculptures has been
conserved using these procedures since 1999. Treated figures
(about 20 by now) exhibit much more polychromy than was
imaginable before. For the first time the terracotta figures ap-
pear as polychrome rather than monochrome sculptures. Guided
by preserved polychrome remnants and archaeological reports,
efforts to understand and reconstruct the original appearance of
individual figures and small units have been intensified since
PAINTING TECHNIQUE AND PAINTING MATERIALS
The terracotta figures were modelled by hand and fired at ap-
proximately 950°C. The reduction firing caused the terracotta
to turn light grey. Basically the figures were fired in one piece,
a remarkable achievement for the time. The quality of the terra-
cotta hardly varies at all; the clay appears to have been prepared
at a centralized station. The potters mastered production of the
1.80 m high figures not only with technical perfection and
expertise, but also with great sculptural skill. The figures are
realistically and expressively modelled down to the details. More
than 80 artists can be identified from the numerous signatures;
some of them were craftsmen from the vicinity who scratched
their names in the clay, others were employees of the imperial
court who used official seals.
Qi lacquer ground
The figures were painted after they were fired. They were first
completely coated with a ground of qi lacquer, onto which the
pigment layer was then applied. The qi lacquer was applied in
two thin layers, the first of which is brown and visibly porous,
perhaps indicating a poorer quality of lacquer. Together both
lacquer layers have a thickness of 30−100 µm. The lacquer did
not penetrate into the terracotta surface: it seems that the terra-
cotta had first been coated with a sealant. Presumably the use
of a sealant made it possible to achieve a smooth surface with
applications of two thin layers of lacquer, without polishing.
This barrier represents one of the conservation problems today,
because it is responsible for the lacquer detaching from the
terracotta surface like a film.
Qi lacquer layers are extremely stable, being sensitive only to
ultraviolet light and changes of humidity. This was already well
known in antiquity and may have been one reason qi lacquer
was used on so many burial objects in Qin Shihuang’s tomb:
not only the terracotta sculptures but also the chariots and the
shafts and textile sheaths of weapons were coated with lacquer.
It might not be merely a historical myth that the complete extinc-
tion of lacquer trees in this region dates back to extraordinary
consumption during Qin Shihuang’s time. Moreover, lacquer was
very valuable. Only the emperor could afford to have everything
he wanted covered with lacquer, an expression of his power and
his endless wealth.
Dust-free rooms and a humid climate are necessary for harden-
ing lacquer. A relative humidity of 65−90% and a temperature of
16−22°C are ideal. These conditions generally exist in Lintong
in May and September; the winters are too cold and dry and the
summers too hot. Probably there were no climate-controlled lac-
quer drying rooms at the time, especially of the necessary size.
Therefore, the production might have been seasonal, with the
terracotta figures being fired in the winter, lacquered in spring
and autumn, and painted during the winter months.
Numerous polychrome layers have been investigated over a 15
year period, but the fragments they came from were chosen for
conservation trials, not for technological questions. Systematic
studies were only started in 2000, using polarised light micro-
scopy (PLM), X-ray diffraction (XRD) and X-ray f luorescence
(XRF). For particular questions scanning electron micros-
copy (SEM-EDX/WDX) and Raman spectroscopy were also
Only inorganic pigments were used for the polychromy. The
range broadly accords with the palette known in antiquity: white
lead, kaolin, yellow and red ochre, cinnabar, malachite, azurite,
and bone black. The extensive use of bone white (i.e. bone ash),
the most frequent white pigment in white and pink layers, is
interesting: it might be explained by the use of animal bones
left over from catering for thousands of workers and artisans. A
violet barium copper silicate was used, nowadays known as ‘Han
purple’ or ‘Chinese pur ple’. This man-made pigment was well
known in Chinese antiquity, but fell out of use after the Han
Dynasty. It was identified for the first time again in 1983
Completely missing, except for golden ochre, is the colour
yellow. There is only one bright yellow layer, hidden under a light
pink one. During the investigations of the project, the colorant of
this layer was identified as vanadinite. This intensive ‘sunflower’
yellow lead-vanadium compound which naturally occurs in lead
ores has not previously been known as a pigment. The same
pigment has also been discovered in a Han Dynasty tomb mural
in Xi’an. Further research on the wider use of this mineral as a
pigment has been carried out, with SEM-EDX , and Raman
and micro-XRF .
One emphasis has been on surveying and interpreting pigment
mixtures and colour nuances, such as the different compositions
and colours used for pink and whitish tones. All flesh colours
were made using mixtures of white and red only. Bone white
was generally employed as the white, sometimes mixed with lead
white. The red pigment was always cinnabar, usually extremely
finely ground (average particle size about 1−3 µm). Pink parts
of the garments were also mixed with bone white, or lead white
and cinnabar, but often the cinnabar is rather coarse. It is still
not clear which of the many detectable variations in the mixtures
are attributable to the work of numerous painters over more than
a decade, and which indicate artistic intent.
Binding media and application techniques
It has not yet been possible to identify the binding medium for
the pigment, which has largely deteriorated during the 2200 years
when the sculptures were covered with moist earth. Aqueous
materials such as animal glue, a plant gum, starch or egg are
likely, whereas oily and resinous binding media do not appear
to have been used in China for sculptural polychromy. The loss
of the binding medium causes many layers to appear completely
opaque. Many of the faces now seem unnaturally pale. The
frequently-found layer structure with a darker pink layer under
a lighter one, must have been part of an artistic concept despite
the considerable thickness of the layers, but it is not visible today
because the layers are completely non-transparent. The ‘fading’,
especially of the faces, often mentioned by Chinese archaeolo-
gists, is simply caused by lightening during the drying process.
The appearance of the still-damp pigment layers on freshly
excavated figures, with slight transparency and bright colours,
is very attractive — and probably closer to the original concept
than the dried-out layers.
The application techniques might provide an indication of
binding media. The pigment layers were generally applied in
a single layer of considerable thickness (0.01−0.8 mm). Still-
visible brush marks indicate experienced and rational application.
Brushes more than 10 cm in width were used for the clothes,
finer ones for more detailed and smaller areas. The fine patterns
on the armour of high ranking officers have extremely precise
lines less than 1 mm in width, which appear slightly raised in
Trials with different kinds of binding media showed that it
is impossible to create layers and surface structures like this
using gums, pure egg or egg yolk, or mixtures with additions of
oily components. Glue and mixtures of glue and egg or casein
gave satisfying results, though even they are difficult to apply
in thick layers.
Surface textures and pastiglia
Visible brushstrokes were perhaps an unavoidable part of the
painting technique, but they were also used as part of the tech-
nique. They follow the natural lines of facial features and they
even seem to indicate thread structure on the garments. This is
particularly evident on the acrobats’ skirts where the brushstroke
runs horizontally on the skirts themselves, but vertically on the
broad white border. In the latter the brushstroke is combined
with fine, raised ornaments. These ornaments of dimensions only
1.2 × 2.5 cm have lines of c.1 mm width with sharp edges, Fig. 1.
The straight brushstroke without a discernible beginning or end
clearly represents the texture of the fabric.
Similar imitations of materials with raised paint are also found
on the bronze chariots and the bronze birds. Tiny dots in stag-
gered rows appear on the white surface of the first charioteer’s
shield and on the disc hanging from his belt, the latter suggesting
a jade bi disc with its typical raised dots, Fig. 2. The feathers on
the white bronze birds are depicted in such a manner that there
was initially speculation as to whether real feathers had been
pressed into the paint. In fact the texture was achieved with
brushes, with the quill of the feather then modelled. All the paint
Fig. 1 Detail of the border of the skirt of acrobat No. 6 (originally white,
but blackened by fire); modelled lozenge pattern (single elements
about 1.2 × 2.5 cm).
Fig. 2 Detail of the charioteer of bronze chariot No. 1, (diameter of disc
layers of this type that were studied are white, thick, and contain
bone white, sometimes also lead white.
Raised decorations as part of white ground layers are known
in Europe under the term pastiglia. In Chinese, the technique
is widespread on Buddhist sculptures, often used in combina-
tion with gilding, and called lifen. On the Qin figures the same
technique is used, but using the paint layer itself. Reconstructions
showed that an effect like this can be convincingly imitated using
bone white in animal glue.
The polychromy exhibits brilliant tones with lively contrasts.
The (probably always) relatively matt colours contrasted with the
glossy lacquer layers. The lacquer not only served as the ground
for the pigments, but was also used as a glossy dark brown colour
to depict hair as well as shoes, armour plates and caps made of
lacquered leather. The areas to be lacquered but not painted were
particularly carefully modelled and smoothed. Colours were used
in bright contrasts next to each other, but mainly each area of
the garment was uniform in colour. Therefore, some collars and
sleeve cuffs which vary in colour over a single area, often from
blue or green to ochre, seemed inexplicable. They were first in-
terpreted as discolourations, but the investigation of structure and
pigments showed this is part of the original concept, reminiscent
The polychromy was not only colourful and individualistic
but also was intended to appear as realistic as possible. The
details and the imitations of various materials contribute to
this effect. The base slabs which were necessary to provide
support for the standing figures were the only unlacquered
components, thus making them practically invisible on the
grey brick floor of the pit. The sculptures seemed to be stand-
ing on their own feet rather than on plinths, a trick intended to
increase the realism of the figures. Further more, originally the
warriors were equipped with weapons from the real Qin Army.
The wooden and textile parts such as the shafts of the weapons,
the covers of the blades and the char iots were also lacquered
The polychromy of one formation
Starting in 1999 it was possible to conserve completely the poly-
chromy on nine figures that were excavated in corridor 18 of pit 2
(which had not been so badly damaged by fire). The sculptures are
part of a unit of kneeling archers who were arranged in a double
row of 2 × 20 figures. Their polychromy can be almost completely
reconstructed: it turned out that all nine archers were painted dif-
ferently. Fig. 3 shows one archer after conservation, and Fig. 4
shows reconstructed colours sketched for all nine archers. There is
no dominant colour within this small troop, and combinations do
not repeat themselves. Archaeologists had already described this
phenomenon in 1988. Military ranks are above all recognisable by
the caps. A few written documents from the Qin Dynasty and the
polychromy of Han Dynasty terracotta armies suggest that there
were no uniforms at the time: soldiers had to provide their own
clothing. The polychrome reconstruction of nine of the archers
suggests how colourful such an army must have looked.
One of the kneeling archers is modelled like all the others but
his face is green. The ‘green face’ is a riddle to the archaeolo-
gists. Close examination as part of the polychromy investiga-
tion showed that there is no colour change (as assumed at the
beginning). Furthermore, not only are his hands and feet flesh-
coloured, but also his neck, the back of his ears and even the
parting in his hair. Though we cannot solve the mystery, at least
we know now that he is not a green-skinned warrior, but a warrior
in green make-up — why, however, no one can say.
Armour with patterns for high-ranking officers
Some figures were not only painted, but had decorations on the
armour with very fine patterns. From the style of their caps, they
can be identified as generals and officers in special positions.
The numbers make it clear that these figures are truly special.
Of approximately 1500 figures excavated so far, only 16 of them
have these particular decorations. There are hardly any pigment
remnants on these figures. Only the archaeologists’ line draw-
ings, a few colour photographs and a few fragments of pigment
layers in the soil (Fig. 5) have survived. None of the patterns
can be interpreted, and neither the repeat nor the colours for
all ornaments were known. In this situation, reconstructing the
polychromy of these figures is like a jigsaw puzzle where slowly
single parts come together, allowing us to reconstruct more and
more of the original polychromy.
Meanwhile, four different types of patterns could be identi-
fied, and the pattern repeats have been completely recorded.
The generals have armour which is edged with geometrically
patterned trimming. On a reddish brown to violet background
black lines form a grid of rhombi which are filled with small
coloured designs. Except for the colours and sequence of the
small designs, this pattern was rather well-documented and thus
easy to reconstruct. On the chest, there is another type of pattern,
Fig. 3 Kneeling archer 002812 from pit No. 2.
but only a few fragments were documented. They were described
showing key- or cucumber-shaped forms which could not be
interpreted immediately. By incorporating all the fragments
and remnants from the figures, it was possible to reconstruct the
pattern in 2005: it obviously shows stylized pairs of birds (prob-
ably peacocks) with suns between them arranged in rows, which
are separated from one another by angular geometric patterns.
These patterns obviously depict woven fabrics. They even imitate
the angular forms and especially the six-sided ‘dots’ which would
be caused by weaving. Even the distortion of the fabric caused
by stretching across the body, is painted: the rhomb-shaped grids
on the trimming are distorted and also the bird pattern is slightly
bent close to the edges.
The angular forms are not a design style. This becomes clear
by examining sculptures with other dresses: the skir ts of the
acrobats have finely curved, not angular, ornaments which seem
to depict embroidered patterns, whereas the white on white pas-
tiglia lozenges on the skir ts borders of their skirts most likely
represent damask or silk gauze with woven-in ornaments. The
artists evidently intended to imitate different types of weaving
techniques for fabric.
Such an exact depiction of fabric raises questions of prototypes
and the significance of the patterns. No fabric from this period
has survived in the vicinity of Xi’an, but several have been pre-
served in graves in east and south China and in Xinjiang. Woven
fabric from Hunan (Zuojiatang near Changsha) and from Hubei
(Mashanling near Jiangling), both dating from the fourth centur y
BC in what was then the southern state of Chu, show stylized
animals with angular dots and edges set off by colours. A piece
of fabric from Mawangdui near Changsha from 168 BC displays
woven peacocks and suns. Silk gauze from the same burial
complex has almost the same ornaments as on the borders of the
acrobats’ skirts. An exact interpretation of the patterns and an
investigation of connections to Central Asia (where peacocks, for
example, originate) via Xinjiang and to southern China (the state
of Chu) have only just begun and will demand more time.
RECONSTRUCTION OF THE POLYCHROMY
Visual impression of the polychromy today
Even the figures that have been conserved are only a pale re-
flection of the terracotta army’s original appearance, because
the polychromy has aged and been diminished. Extensive gaps,
discolouration to yellow from the yellow soil, as well as lost sur-
face textures and thin top layers impair the effect. The missing
Fig. 4 Reconstruction of the nine kneeling archers from pit no. 2,
excavation area 21, corridor 18.
Fig. 5 B-0101, the only fragment of the armour of general T9:1 from
pit No. 2, mounted in plaster (longest dimensions 13 × 15 cm).
binding medium makes the pigment layers seem too light and
opaque. Consolidation measures also have the unavoidable effect
of changing saturation and depth of colour. This could be a step
back towards the original appearance, but as it still impossible to
say how much saturation and gloss the pigmented layers had, it
can also be a falsification of the visual impression. As the colours
of the lacquer and the terracotta surface are also slightly changed
by both consolidation methods, the conserved fragments actually
might be quite far away from the original impression.
The dust deposits in Lintong are another unresolved problem
because it is not possible to remove them from the mechanically
sensitive pigment layers. Moreover the figures’ ‘accessories’ are
gone: only the bronze components of the wooden accessories are
preserved; the weapons and quivers are missing. In view of the
progressive drying out of the pits it is unlikely we will encounter
better preserved polychromy.
Depicting the results of the investigations
Descriptions, line drawings and data on the colours record infor-
mation but do not convey a visual impression, especially in regard
to colour nuances, surface textures and patterns. Photographs
only reflect the diminished state of the figures today (Fig. 6), so
that details and colours often can hardly be imagined. Therefore
attempts were made to visualize existing findings using colour
sketches, gradually moving from small coloured fragments
(Fig. 7) back to the original polychromy and the overall design.
It was not possible to reconstruct the entire polychromy on any
one of the figures with patterns. But the prevailing analogies of
the basic colour schemes and the arrangement of patterns within
one type of figures allow a relatively assured completion of the
design, Fig. 8. Colour sketches have the disadvantage that the
modelling of three-dimensional figures, and surface characteris-
tics such as br ushstrokes and gloss, are difficult to represent.
These problems can partly be solved by computer anima-
tion combining 3-D scans, photographs and the application of
virtually designed surfaces which can be used to reduce dis-
colouration and fill in missing parts. Even several figures can be
combined and shown in a reconstruction of the original surround-
ings inside the pit. Computer animation of the polychromy has
been par t of the ongoing work of the project since 2003.
Reconstruction on replicas
Another method consists in colour reconstruction on replicas
of the sculptures. The same materials and techniques as used
in antiquity can be used, and the figures can be experienced in
space like the originals. The possibilities and limitations of the
materials themselves, such as the influence of binding media on
Fig. 6 The back of the general T9.
Fig. 7 Reconstructed colour sketch of B-0101.
Fig. 8 Reconstructed colour sketch of the back of general T9:1 from pit
colour depth, surface texture and gloss, provide the constraints
in this approach. Moreover, the issue arises of how to deal with
gaps for which there is no information; other than on a drawing,
they cannot simply be left blank.
Despite these limitations and problems, an attempt to make
polychrome reconstructions has been started on two replicas in
January 2006. The replicas are made of gypsum, remodelled by
hand and coated with raw qi lacquer, which results in a shiny,
black-brown surface. The pigments were chosen in accordance
with the paint analyses and have been acquired from Beijing.
‘Chinese violet’, which is no longer available today, was made
expressly for this purpose.1 Various binding media such as animal
glues, egg white, egg yolk, gums, starch and casein were tested in
order to achieve an effect similar to the originals. Parallel to this,
attempts were again made to identify the binding medium, and
for the first time, there might be a result: preliminary analyses
seem to indicate the presence of egg . Therefore, a mixture of
egg with skin glue has been used as the binding medium for now.
Even if this reconstruction remains speculative in some respects,
it might give an initial impression of the original appearance of
a terracotta warrior.
Research results give an idea of the complicated polychromy
of the terracotta figures. The extensively used valuable pig-
ments and qi lacquer were both expensive materials that only an
emperor could afford. Through its paint scheme the terracotta
army became a highly realistic and colourful replica of a Qin
Dynasty army. Painted and modelled patterns imitate contempo-
rary fabrics in a realistic manner and further enrich the detailed
If the paint layers and organic materials had survived, one
would see a structurally organized but finely detailed, colour-
ful crowd instead of a grey army. Rows of figures would be
broken up by protruding weapons, quivers and chariots, whose
wooden components were also glossily lacquered and painted.
It is not possible to imagine this effect today. Conservation and
reconstructions based on research on all available material can
only give an idea what the polychromy might have looked like
on single, isolated figures. Nevertheless, it might be a first step
to understand the original appearance of the terracotta figures
found as burial objects around Qin Shihuang’s grave.
However, it can be assumed that the terracotta army was never
to be seen in antiquity as it is today. The height of 3.20 m of
the dividing walls would hardly have allowed anyone to see the
figures. Moreover the corridors were certainly never open, since
the pits were not roofed over by halls. Rather one must imagine
that the figures disappeared piece by piece, according to an
established plan, into the darkness of a corridor which was
covered with wooden beams and locked against thieves at its
entrances. Thus not even the successor of Qin Shihuang would
ever have seen Qin Shihuang’s terracotta army in its complete
This work would not have been possible without intensive collaboration
within and beyond this project: Dr Daniela Bathelt, Prof H. Berke, Rong
Bo, Prof Lin Chunmei, Dr Patrick Dietemann, Wang Dongfeng, Akram
El Jarad, Dr Gerd Guelker, Dr Susanne Greiff, Dr Christoph Herm, Felix
Horn, Mrs Fan Juan, Dr Herbert Juling, Dr Arne Kraft, Wang Liang,
Klaus Rapp, Dr Ingo Rogner, Carolin Roth, Dr Soong Shing-Müller,
Zhou Tie, Dr Cristina Thieme, Vojislaw Tucic, Dr Lucien van Valen,
Prof Yuan Zhongyi, Liu Zhancheng, and Dr Bettina Zorn all contributed
to the work.
1 Yang Hsien-yi, Selections from Records of the Historian by Szu ma
Chien, Beijing, 1979 (unpublished).
2 Watson, B., Records of the Grand Historian by Sima Qian, Columbia
Press, Hong Kong (1993).
3 Nienhauser, W. (ed.), The Grand Scribe’s Records. Vol. 1 The Basic
Annals of Pre-Han China by Ssu-ma Ch’ien, Indiana University
Press, Indianapolis (1994).
4 Rawson, J., (ed.), The British Museum Book of Chinese Art, New
York (1993) 138−139.
5 Ledderose, L., ‘The Magic Army of the First Emperor’, in The
Terracotta Army of the First Chinese Emperor Qin Shihuang, ed.
C. Blänsdorf, E. Emmerling and M. Petzet, Munich (2001)
6 Emmerli ng, E, Thieme C., Zhou Tie and Zhang Zhijun, ‘Initial
conservation work 1991−1995’ in The Terracotta Army of the First
Chinese Emperor Qin Shihuang, ed. C. Blänsdorf, E. Emmerling
and M. Petzet, Munich (2001) 536−537.
7 Simon, S., ‘Conservation 1995 — Test Series and Quality Control’,
in The Terracotta Army of the First Chinese Emperor Qin Shihuang,
ed. C. Blänsdorf, E. Emmerling and M. Petzet, Munich (2001)
8 Herm, C., Zhou Tie, Th ieme, C., and He Fan, ‘Results of Conser-
vation Test Series 1996/97’, in The Terracotta Army of the First
Chinese Emperor Qin Shihuang, ed. C. Blänsdorf, E. Emmerling
and M. Petzet, Munich (2001) 584−593.
9 Rogner, I., Festigung un d Erhaltung der polychromen Qi-
Lackschichten der Terrakottakrieger des Qin Shihu angdi durch Be-
handlung mit Methacryl-Monomeren und Elekt ronenbestrahlung,
unpublished dissertation thesis, Munich (200 0).
10 Bathelt, D., Farbigkeit in der Antike — Ent wicklung chemischer
Methoden zur Erhaltung der Farbfassung der Terrakottaarmee des
ersten chinesischen Ka isers Qin Shihuangdi, unpublished disser ta-
tion thesis, Munich (2005).
11 Testing and Optimising of Conservation Technologies for the Pres-
ervation of Cultural Heritage of the Shaanxi Province, Bavar ian
State Depa rtment for Historical Monuments Munich (1990−2004).
12 Fitzhugh, E.W., and Zycherman, L., ‘An early man-made blue pig-
ment from China — barium copper silicate’, Studies in Conservation
28 (1983) 15−23.
13 Fitzhugh, E.W., and Zycherman, L., ‘A purple ba rium copper sili-
cate pigment from early China’, Studies in Conserva tion 37 (1992)
14 Rapp, K., Geological Institute of Ludwig Maximilians University
Munich, personal com munication, August 2003 and November
15 Grieff, S., Roman Germanic Central Museum, personal communica-
tion, November 2004.
16 Dietemann, P., personal communication, November 2004.
Catharina Blänsdorf has specialized on the conservation of wooden
sculpture and panel painting. She received her MA diploma in conserva-
tion from the Academy of Arts in Berne, Switzerland. She has worked
on the ‘China Project’ in the Bavarian State Department for Historical
Monuments (Bayerisches Landesamt für Denkmalpflege) in Munich since
1998, concentrating mainly on conservation of the polychromy of the
Terracotta Army of Qin Shihuang. Address: Bayerisches Landesamt für
Denkmalpflege, China Project, Hofgraben 4, D-80539 Munich, Germany.
Xia Yin studied chemistry at North-West University in Xi’an. He has
worked in the conservation department of the Museum of the Terracotta
Army since 1997, participating in the conservation of polychrome ter-
racotta sculptures and stone armour. He has focused on the history and
identification of pigments, analysing samples from many sites in China.
Address: Xia Yin, Museum of the Terracotta Warriors and Horses of Qin
Shihuang, Laboratory, 710600 Lintong, Shaanxi Province, PR China.
Prof H. Berke, University of Zurich, is conducting intensive research on
Chinese blue and violet. He has provided 1 kg of violet to the project for
painting the replicas.