Unveil the gold – Revealing metal threads and decorative materials of early 20th 
century traditional Chinese children’s hats  
 
Angela Cheung1, Caroline Solazzo2, Wai-shan Tsui1 
 
1 Conservation Office, Leisure and Cultural Services Department (LCSD), the 
Government of the Hong Kong Special Administrative Region of the People’s Republic 
of China (HKSAR) 
2 Museum Conservation Institute (MCI), Smithsonian Institution, United States of 
America  
 
Keyword: Chinese children’s hat; metal threads; membrane threads; proteomics, leather, 
fish glue, yellow croaker, SEM-EDX 
 
 
Abstract 
This paper presents the study of metal threads and other decorative materials on eleven 
traditional Chinese children’s hats dated to the early 20th century. The lavishly decorated 
hats are presented to children to protect them from evil and bless them for good health 
and fortune. Those hats were selected for display in the 2015 exhibition Wearable 
Blessings: Traditional Chinese Children’s Clothing at the Hong Kong Heritage Museum 
(HKHM).   
In addition to the conservation work, the hats were fully studied to encompass analysis of 
the materials and understand the craftsmanship of the exquisitely decorated hats.  In 
particular, various forms of metal threads were identified. The majority of the threads 
were revealed to be “membrane threads”, mostly gold or silver lamellae on leather 
substrates and are used in the hats as flat strips or foils. Among the eleven hats, nine are 
also decorated with metal-wrapped threads with fibrous cores (silk, linen or cotton).  The 
base substrate of the membrane thread was identified as either sheep or goat skin by 
means of proteomics in eleven samples tested from six hats.  These findings coincide 
with the information recorded in an ancient Chinese encyclopaedia “Chinese Technology 
in the Seventeenth Century”. Using proteomics, we also found a collagen signal matching 
to a fish species, Larimichthys crocea or large yellow croaker that is suspected to come 
from the glue binding the metal onto the base substrate. Other than the metal threads, 
feathers were also identified by microscopic examination on one of the hats. In terms of 
craftsmanship, leather gold embroideries and the hidden framework of a three-
dimensional lotus ornament were examined and revealed by X-radiography respectively.  
This study demonstrates that the integration of scientific analysis with conservation is 
valuable in the investigation of material cultures and intricate craftsmanship of the 
historical objects.  
 
Introduction 
1 
 
The study of eleven traditional Chinese children’s hats was initiated in 2015 to 
understand more about the materials and craftsmanship used in their production as, to the 
best of our knowledge, few studies have been published on this topic (Luo 2018). Those 
hats are richly decorated with metal threads and were showcased in an exhibition 
“Wearable Blessings: Traditional Chinese Children’s Clothing” held in the Hong Kong 
Heritage Museum (HKHM).   
The exhibition displayed the beauty and exquisiteness of Chinese children’s clothing 
dated from the 19th to early 20th century.  The clothing included hats, clothes, dudou (an 
undergarment halter covering the chest and belly), waistcoats, shoes and boots, earmuffs, 
locks, etc. The decorative symbolic motifs on the clothing are not only visually appealing 
but also shed light on the traditional Chinese culture that embraces blessings for children. 
The motifs present blessings for longevity, happiness, success, abundant offspring, etc. 
The incorporation of precious metals to textiles dates back several thousand years.  The 
earliest reference may probably be the Bible’s Old Testament, Exodus 39:2-3, which 
describes the making of Aaron's vestment for service during the 12th -13th  centuries BC 
(Járó 1990). Membrane thread made of metal coated leather, animal gut or paper in strip 
or wound around a fibrous core appeared in the 11th century (Járó 1990; Karatzani 2012).  
The gold membrane threads, the so called “Cyprus gold threads” or “gold saving” 
threads, began to spread to Europe from Byzantium or possibly from the Western Asian 
regions. They were lighter and more flexible than solid metal threads.  This reduced the 
weight of the clothing and the production cost (Hacke et al. 2004; Járó 1990). 
The use of precious metals on textiles in China also has a long history.  The application 
of metal threads, in particular the gold-wrapped threads, could be traced back to the 
earliest archaeological fragments dated to the 3rd-4th centuries, which were excavated in 
Xinjiang (Dang et al. 2014; Hu and Yu 2016; Lu et al. 2017).   The production of a gold 
brocade robe was mentioned in a historical text, Suishu “The book of Sui dynasty (AD 
581-618)”, the biographic chapter of He Chou.  He Chou was an imperial officer and was 
requested by the emperor to make a gold brocade robe of exquisitely elaborate design 
with reference to the one presented by Persia  (Zhao 2005).   
A unique textile collection embellished with a vast amount of metal threads of the later 
Tang dynasty (AD 618-907) was unearthed in 1987.  The archaeological textiles were 
excavated at the Famensi (Famen temple), a Buddhist pagoda dating back to the 6th 
century at Fufeng, a county in Shaanxi Province.  The excavated textiles comprised a lot 
of metal threads, most of them were solid gold strips wound around a fibrous core 
( Karatzani et al.  2009; Yang et al. 2013).  One silver-wrapped thread in a piece of cloth 
was reported to likely be made with an organic paper support (Lu 2015).   Another 
important group of Tang dynasty’s textiles illustrating the presence of metal threads are 
the textiles from Dunhuang.  Gold strips backed with papers were thought to be used on 
silk tapestries (kesi) found at the library cave of Qian Fo Dong.  (Zhao 2005; British 
Museum 2020 a).  Moreover, one piece of embroidered fragment was reported to have 
the outlines of motifs originally made in gold strips.  Most of these strips have 
2 
 
disappeared, revealing a black substrate, probably from an animal tissue (Wang and Zhao 
2012; British Museum 2020 b).  Animal skin was notable as the backing of gold threads 
used in Nashishi (also known as Nachisi), the most famous gold woven brocade in the 
Yuan dynasty (AD 1271-1368). Nachisi is the transliteration of Nasich, the Persian word 
for gold-wefted brocade which comes from the Arabic language (Zhao 2005 and 2012). It 
was described in a historical text “Dao yuan xue gu lu”, written by Yu Ji who served at 
the Yuan imperial court, as woven with gold threads attached onto leathers (Mao 2020; 
Zhao 2005). 
Additionally, the making of metal coated skin strip (or generally-speaking membrane 
thread), also known as “leather gold”, was recorded as early as 1637 (Ming dynasty) in an 
encyclopaedia “Chinese Technology in the Seventeenth Century”: “In Shensi (Shaanxi 
province) the people make “leather gold” by affixing gold leaves to sheep skin curved to 
an extreme thinness, so that it can be cut and made into articles of clothing or ornament, 
all of which gleam with gold splendor” (Sung 1966). In China, a single Chinese character 
“羊” (yang) is the collective description for all breeds of Caprinae family which includes 
both “sheep” and “goat”.  It was noted in the encyclopaedia that there are two kinds of 
sheep: “One is known as woolly sheep……  The other kind of sheep is called yu-tiao 
sheep (this is a foreign term)….. the Shensi people called this animal [cashmere] goat 
( Sung  1966).”  
In connection to the analysis and characterisation of metal threads, a large literature has 
been published.  Optical microscopy is frequently used as a preliminary examination tool 
to investigate the type of metal threads, the deterioration of the surfaces, nature of core 
fibres, substrate materials, etc (Járó and Tóth 1991).   Scanning electron microscopy 
(SEM) can be used in conjunction for the surface morphology investigation. Moreover, 
elemental compositions can be deduced using SEM coupled with X-ray energy dispersive 
spectroscopy (EDX) i.e. SEM/EDX (Kohara et al. 1998; Nord and Tronner 2000;  Járó et 
al. 2000; Hacke et al. 2004; Muros et al. 2007, Rezić et al. 2010; Oraltay and Karadag 
2020).  Recently, various electron microscope techniques and micro-Raman spectroscopy 
have been used to study the production technologies of medieval metal threads. 
(Weiszburg et al. 2017).  Imaging analytical techniques including micro-X-ray 
fluorescence (XRF) scanning and micro-X-ray radiography have been used on 
archaeological metal threads as no samples are required for the investigation. (Hložek et 
al. 2019).  
In this study, decorative elements were firstly examined by visual examination such as 
optical microscopy and scanning electron microscopy (SEM) to obtain the surface 
morphologies, followed by SEM/EDX analysis. This initial examination found that nine 
hats were decorated with metal-wrapped threads.  All had decorations of flat metal strips 
or foils. Close examination of the metal strips and foils revealed that the metal likely 
adhered to a substrate made of animal skin.  While the analytical studies on metal threads 
are well established for the metal composition and fabrication techniques, less work has 
been reported on the identification of organic substrates of membrane metal threads 
3 
 
(Indicator et al. 1989). Yet, the identification of leather is commonly conducted by 
microscopy (Haines 2006) and infrared spectroscopy for the morphological comparison 
of the grain structures and possible differentiation of the natural and synthetic leathers 
respectively (Shao 2005; Luo et al.  2011).  Nevertheless, the identification may pose 
difficulties as the characteristic appearances of animal skin leathers may be obliterated by 
tanning, leather treatment and finishing, or deterioration.   DNA analysis from leather 
may be an option but DNA may be damaged during the manufacturing process such as 
tanning (Merheb et al.  2014 ; Matar and Merheb 2016).    
Proteomics has been increasingly used in cultural heritage studies since early 2000 to 
identify protein based materials such as protein based binders (Tokarski et al. 2006; Leo 
et al. 2009; Ma et al. 2017), adhesives (Dallongeville et al.  2011; Dallongeville et al.  
2013; Zhu et al. 2019) in artworks and polychromies as well as archaeological objects 
(Kumazawa et al. 2016; Vinciguerra et al. 2016; Solazzo et al. 2016).  Proteomics is the 
study of the protein composition of tissues or other biological systems using an array of 
techniques based on mass spectrometric analysis. In relation to its application to historical 
textiles and clothing, including metal threads, the characterisation of ancient proteins can 
increase our knowledge of techniques of fabrication, and material availability, variety, 
or trade (Solazzo et al. 2013; Solazzo 2019). In the study of metal threads made with an 
organic support layer, proteomics is used for multiple purposes:  
1. To characterise the type of support tissue. Although microscopy analysis should be 
sufficient to differentiate metal threads made with an animal membrane (usually 
intestinal tissues) or leather (made from skin), the thickness (e.g. thin skin from a 
young animal) or deterioration of the support layer might raise uncertainty in some 
cases. While leather is dominated by collagen, intestinal membranes have, in 
addition to collagen, an array of other proteins such as actin and myosin (Popowich 
et al. 2018). These proteins can be used to differentiate the two types of collagenous 
tissues. 
2. To identify the source species of the support layer. In recent proteomics studies, the 
source is often from a bovidae species (sheep, goat, cow) (Popowich et al. 2018; 
Solazzo et al. 2019; Scibè et al. 2020). 
3. To determine the presence, nature and source species of the adhesive used to bind the 
metal to the organic substrate. This adhesive can be, for example, collagen glue or 
egg-based (Solazzo et al. 2019; Scibè et al. 2020).    
  
Proteomic was thus conducted further on eleven samples from six hats to identify the 
animal source of the skin. It is a state-of-the-art tool to study membrane threads and only 
minute samples (less than 0.1 mg) are required. Species identification of animal skins and 
the binding adhesive of the membrane threads can provide insight about the materials and 
craftsmanship of the decorative elements.  The scientific findings may possibly link up 
the material knowledge of the membrane threads with the traditional production 
techniques which are rarely documented. The overall study offers added-value benefits to 
4 
 
the long term preservation of the traditional Chinese children’s hats of the early 20th 
century. 
 
Traditional Chinese children’s hats 
Blessing 
The birth of a child, especially a baby boy, is a great event for Chinese families. In rural 
China, the mortality rate of children was high before the mid-20th century due to 
inadequate hygiene and less favourable medical condition. (Leung 2010; Sun and Zhang 
2008). Mothers or relatives conveyed their blessings to the children by hand-stitching the 
hats that were embellished with symbolic and auspicious motifs. Apart from protecting 
the children from cold, hats worn by children denoted wealth, luck and longevity. There 
was a wide variety of children’s headwear, some were designed to be worn on festivals 
and special occasions.  Children wore different types of hats from infancy to early teens; 
those hats intended to ward off evil spirits when the wearers were young and transitioned 
to bring future success to them when they grew older (Garrett 2008; Sezto and Garrett 
1990).   Some hats were especially designed for boys such as animal hats or scholar hats.  
Ferocious animal hats such as a lion’s head or tiger’s head (Figures 1a-e) were frequently 
depicted for protection. A back flap was added to cover the back of the neck in cold 
weather (Figure 1aii). Scholar hats were imbued with good fortune for future official 
examinations.  Shaped like the ones worn by scholar officials, they were intended to give 
boys a head start on a future career as high-ranking officials. In addition, there were hats 
in the shape of a melon or pomegranate or lotus which symbolised endless descendants.  
(HKHM 2015; Leung 2010).  A brief description of the hats in this study is presented in 
Table 1 and their images are shown in Figures 1a-k.   
Children hats are heavily decorated with a variety of decorative motifs such as animals, 
plants, flowers, Chinese calligraphy, legends, etc (Garrett 2008; Lu and Cui 2016). These 
motifs have their implied auspicious meanings.  The “meaning” is presented indirectly 
through three main forms of representation: symbolism (usually of the object’s form, 
colour and nature), homophones and abstract motifs (Leung 2010). Animals are 
understood to bestow unique abilities such as strength and power, and the children are 
protected by virtue of these same qualities.  Ferocious animals like tigers and lions are 
believed to protect them from harm. 
The powerful tiger is a divine creature in Chinese folklore. Parents, especially in North 
China, believe that tigers frighten away the evil spirits and ensure a prospective future for 
their children (Leung 2010; Sezto and Garrett 1990).  Wearing a tiger hat can ward off 
evil spirits or fool them into thinking that the child is a fearsome animal (Sezto and 
Garrett 1990).  If the tiger hat was worn on birthdays or special occasions, it also brought 
luck and peace. The colour of a tiger hat might indicate its place of origin. Hats in natural 
yellow colour of the tiger are common in Shanxi, Shaanxi, Hebei, and Shandong (Cao 
2010), while black tiger hats might originate from the northeast provinces of 
Heilongjiang, Jilin, and Liaoning, where the costumes were more unconstrained in style 
(HKHM 2015).  
5 
 
A typical feature of tiger hats is the presence of the Chinese character “王” (pronounced 
as “wang”), literally meaning “King” on its forehead between the eyes. The eyes and ears 
on a tiger hat are believed to spot danger and hear the approach of evil. Its mouth displays 
a full set of bared teeth, and the wavy embroidered stitches denote the tiger’s stripes. A 
second animal’s face on top of the tiger is believed to provide even greater protection 
(Figure 1f) (Garrett 2008).  
In the Chinese language there are more written characters than spoken words. Therefore, 
many Chinese characters share the same pronunciations, i.e. homonyms with embedded 
symbolic meanings (Sung 1979).  Two examples of homonyms are fish and bat. The 
Chinese character for fish (yu魚) is pronounced the same as the Chinese character for 
“abundance” or “surplus” (yu餘). Therefore, fish is a symbol for “more” in the sense of 
“more” good luck and good fortune. The Chinese character for bat (fu蝠) is a pun on 
“fortune” (fu福). If the bat is shown flying upside down i.e. “inverted fu” (fu dao蝠倒), 
bat and fortune together form the rebus of “fortune arriving” (fu dao福到) (Sezto and 
Garrett 1990; Sung 1979).  
Flowers are common motifs on the hats, and depict wealth, honour, and good fortune 
(Sezto and Garrett 1990).   Lotus, in particular, is a common motif on those children’s 
hats under study.  Lotus is pronounced as “he hua” (荷花) or “lian hua” (蓮花)  in 
Chinese,  the Chinese character  “lian” (蓮/連)   has the same meaning as to bind or 
connect (in marriage) and “he” (和) is a homophone for harmony.  Lotus seeds are called 
“lian zi” (蓮子), the Chinese character and pronunciation for “zi” (子) is same as child or 
son.   While plants usually produce flowers before bearing seeds, the lotus produces 
flowers and ripened seeds at the same time; as such the lotus is interpreted as the symbol 
of “soon giving birth to a noble son” (Sung 1979; HKHM 2015).  One of the lotus hats is 
decorated with a large lotus seedpod in the centre (Figure 1k).  The crown of the hat 
features a three-dimensional lotus with a figurine boy emerging from it. The boy is 
holding a seedpod studded with lotus seeds. These motifs convey the wishes for a son to 
be born every year and for the family to be succeeded with sons and grandsons.  Another 
type of fruit, longan is also noted on one of the tiger hat (Figure 1 aii).  Longan is also 
called guiyuan(桂圓) (the dried form). “Gui” (桂/貴) is a pun on “plucking the 
osmanthus”, which alludes to “obtaining high honours at the imperial examination”, 
“yuan” (圓 /元) is a pun on “supreme ultimate”, suggesting “achieving the highest scores 
in three consecutive examinations” (HKHM 2015).   
Table 1. Description and decorative features of children’s hats (dated 1900 -1930) under analysis  
Item Object  Features Ground fabrics Figure(s) 
no 
1 Orange tiger- -embroidered with the fruit Orange colour- 1ai & 
head hat longan (guiyuan) and the silk 1aii 
6 
 
mythical Chinese unicorn qilin 
motifs on neck flap 
-decorated with metal-wrapped 
threads as the eyebrows 
- embroidered with gold foils 
on neck flap 
2 Yellow tiger- -embroidered with floral Yellow colour- 1b 
head hat  motifs on metal foils cotton 
- decorated with metal-
wrapped threads as the 
eyebrows 
-decorated with metal-wrapped 
threads on neck flap 
3 Black tiger- -embroidered with floral and Black colour- 1c 
head hat lotus motifs  silk 
- decorated with metal-
wrapped threads as the 
eyebrows 
-decorated with brown strips  
4 Black tiger- -embroidered with the Chinese Black colour - 1d 
head hat mythical bird phoenixes, silk 
flowers , red and blue women 
and lotus 
-decorated with metal-wrapped 
threads as the eyebrows 
-decorated with embroidered 
gold foils  
5 Double-faced -superimposed by a yellow Orange colour – 1e 
tiger-head hat tiger’s head cotton  
-decorated with gold longevity Black colour -
roundels and fu福 (fortune), lu silk 
祿(emoluments), shou壽 
(longevity) Chinese characters 
  
6 Dragon-head -embroidered with floral Green colour - 1f 
hat motifs  silk 
-decorated with gold strips 
7 Hat with -decorated with lotus motifs Blue colour - 1g  
double-fish and gold strips on the goldfish  silk 
and lotus -decorated with green feathers 
motifs in the mouths of goldfish as 
waterweeds 
-decorated with metal-wrapped 
threads 
7 
 
8 Hat with -decorated with gold strips on Black colour- 1h 
double-fish a pair of catfish on the two silk 
and lotus sides and the lotus motifs at 
motifs the front  
-decorated with metal-wrapped 
threads 
9 Hat with - shaped like a scholar hat Black colour - 1i 
auspicious -embroidered with numerous silk 
motifs auspicious motifs such as 
inverted bats 
- decorated with gold strips 
and metal-wrapped threads 
 
10 Hat with bat, -decorated with metal-wrapped Black colour - 1j 
peach and threads silk 
ruyi motifs  -embroidered with motifs such 
as  ruyi (wishes fulfilled), bats, 
etc  
-decorated with a few small 
brown strips around the eyes, 
nose and mouth of a bat 
11 Hat with boy -embroidered with lotus motifs  Black colour - 1k 
and lotus -decorated with a three- silk 
motifs dimensional (3D) lotus 
ornament with a boy figurine 
inside the lotus   
-decorated with metal-wrapped 
threads on the 3D lotus petals    
-decorated with brown strips  
 
ai
i  
  
    
8 
 
b
aii 
c
a 
 
f     g 
a 
 
  
  
k 
  
 
 
  
Figures 1a-k.  Images of the children’s hats under study. Photos ©Hong Kong Heritage 
Museum 
 
 
Elements of decoration and craftsmanship 
The children’s hats under study are elaborately decorated with appliqué, embroideries, 
metal threads, embroidered metal foils and other decorative items such as feathers.   The 
feathers represent waterweeds in the mouths of two facing goldfish of a hat (object no 7, 
Figure 1g).  The most commonly used techniques for decorating children’s hats are 
embroidery and appliqué (Figures 2a and b). Stitches include chain stitch, satin stitch, 
couching (a way to anchor gold and silver thread), and dazi (knot stitch or “forbidden” 
stitch) (Lin 2007). The stitches and threads are the deepest expressions of a mother’s 
affection to her child.   The auspicious motifs include longan (guiyuan), lotus blossoms 
and seedpods, phoenixes, peaches, ceremonial scepter ruyi, etc. 
 
9 
 
  
Figure 2.  Embroideries on the children’s hats (a: object no. 6, b: object no. 7).  Microscope 
photos ©Hong Kong Heritage Museum  
 
In addition to embroideries, metal threads are extensively used on the children’s hats.  
The metal-wrapped threads are frequently found as the eyebrows of the tiger hats. They 
are also adopted as the outlines (Figure 2b) or decorative components of the 
embroideries.  Apart from the metal-wrapped threads, the abundant metallic decorative 
elements on the children’s hats are in the form of gold or brown strips (foils) and they are 
flexible in nature.  Some golden colour foils combine with embroideries to form an array 
of motifs. Nevertheless, there is limited information about the manufacture techniques of 
the embroidered “gold” motifs that are stitched on the children’s hats.  
The constructing technique of embroidered golden foils is hence an area of interest to be 
explored. X-radiography is a non-destructive and non-invasive imaging technique.  The X-
radiography of textile objects can reveal construction, manufacturing techniques, use, 
patterns and repair details which are invisible to the naked eye ( Brooks and O'Connor 
2007; Hackett 2011). Some embroidery threads of two women figures of a tiger hat 
(object no. 4) were lost and the papers underneath were exposed i.e. the embroideries 
were stitched on top of paper supports.  A tiny piece of gold emerged at an area where a 
small area of paper was lost. Gold foils may possibly be the underlining layers of the 
papers that were not only decorated along the borders of the embroideries. Therefore, the 
torn areas of two embroidered women figures with gold foils were X-rayed to seek for the 
layering information.  
 
Conservation  
The major conservation problem of the Chinese children’s hats was deformation. Those 
hats might not have been properly stored before they entered the museum collection.  The 
hats were seriously deformed or disfigured. For instance, a three-dimensional (3D) lotus 
ornament with a boy figurine featuring at the top of a hat (object no. 11) suffered 
deformation (Figure 3a) which was possibly caused by crushing.  The boy figurine is 
surrounded with the flattened petals of the lotus and fastened to the petals by stitching.   
Some metal wires are exposed on the underside of the hat where the linings were 
10 
 
damaged.   X-radiography was conducted to study the possible unseen affixing 
mechanism of the lotus ornament and the boy figurine to aid in the treatment decision for 
realignment. Finally, the respective lotus and the boy figurine were straightened back to 
upright positions by manipulating the internal wirework mechanically and re-shaping the 
lotus petals by humidification (Figure 3b).   
Generally, basic surface cleaning was done with fine brushes using vacuum cleaner with 
adjustable suction to the children’s hats.  Afterwards, they were reshaped by 
humidification and the internal shapes were maintained with the support of tailor-made 
cushioning materials, made from polyethylene foams or Fosshape®. Fosshape® is a non-
woven polyester felt-like materials which can be activated by dry or steam heat to mould 
in the desired shape and regularly used in textile conservation.  Due to the presence of 
metal threads and leather substrates, localised controlled humidification by using 
ultrasonic humidifier or moisture permeable material Gore-tex® was performed. Lastly, 
some tiny metal membrane strips were flaked slightly and Lascaux was used for 
consolidation.   
 
 
 
 
Figure 3. (a) Deformed 3D lotus ornament with a boy figurine featuring at the top of a hat (object 
no. 11); (b) The 3D lotus ornament with a boy figurine on top of the hat was re-positioned. Photos 
©Hong Kong Heritage Museum 
 
Experimental   
Sampling  
Most of the samples were taken from the loose or partially detached pieces from the 
children’s hats for SEM/EDX and proteomic analysis. A small portion of the samples 
11 
 
were taken from the tiny torn or unobtrusive areas.  As such, only selective metal threads 
on the children’s hats were analysed.  
 
Microscopic and SEM/EDX examination 
Optical microscopy, a Leica M80 stereomicroscope was firstly used to investigate the 
types of metal threads on eleven hats. Samples were taken for subsequent examinations 
with a Leica MZ60 stereomicroscope to reveal the morphologies and a FEI Quanta 200 
SEM/EDX to obtain highly magnified images as well as elemental compositions.  Nine 
hats display the presence of metal-wrapped threads i.e. metal wound around a fibrous 
core.   Samples of metal-wrapped threads from six hats were taken for SEM/EDX 
analysis (Tables 1 and 2: Object no. 1, 2, 4, 7, 10 and 11).   Separately, metal strips or 
foils (gold or brown foils) from ten hats (Tables 1 and 3:  Object no. 1 to 9 and 11) were 
taken for SEM/EDX analysis.    Cross section of metal strips or foils were also examined 
using a FEI Quanta 200 SEM/EDX and an Olympus BX60 polarizing microscope.  A 
green feather-like material representing the waterweeds in the mouths of two goldfish of 
a hat (object no. 7; Figure 1g) was also examined.      
X-radiography 
 
Embroidered golden foils on object no. 4:  
X-radiography was used to reveal if the golden foil embroideries were supported wholly 
with “leather gold” beneath or just decorated along the borders of the embroidered 
motifs.  Two areas of slightly torn embroidered women figures with gold foils (object no. 
4) were X-rayed by a Faxitron®, X-ray System model 43855C, at an accelerating voltage 
of 20 kV and exposure time of 2 sec.   The penetration of X-ray radiation is dependent on 
its energy, the type, and thickness of the object to be analysed. Organic materials such as 
textiles and papers are of low atomic numbers, the use of low energy or soft X-ray is 
known to be effective in producing detailed radiographic images (Brooks and O'Connor 
2007).  Low X-ray energy at 20kV was selected to obtain high contrast images.   
 
A three-dimensional (3D) lotus ornament with a boy figurine on object no. 11: 
X-radiography was conducted to study the possible affixing mechanism of a deformed 
lotus ornament (object no. 11) to aid in conservation treatment.  The image was taken by 
Faxitron®, X-ray System model 43855C, at an accelerating voltage of 120 kV and 
exposure time of 30 sec.  High energy X-ray was used to obtain a contrast image of the 
hidden metal wire structure inside layers of fabrics.    
 
Proteomics 
 
Eleven samples from six hats were analysed: Object no. 9, 8, 11, 7, 2, 6. The 
experimental protocol was adapted from previous studies (Solazzo et al. 2013):  
 
12 
 
Extraction:  
Proteins were extracted from samples of about 0.1 mg or less by solubilization of the 
leather strips in 100 µL of a solution of 8M urea, 50 mM Tris and 50 mM tris(2-
carboxyethyl)phosphine (TCEP) at pH 8.0, left overnight in shaker at maximum speed. 
The whole supernatant was alkylated for 45 min in the dark with 10 µL of 400 mM of 
iodoacetamide for a final concentration of 40 mM. Samples were dialyzed for 6h in 
dialysis units at 3500 Da (Thermo Scientific™ Slide-A-Lyzer™ 3.5K MWCO MINI 
Dialysis) with 100 mM Ambic (ammonium bicarbonate) at pH 8.0 (two changes). 
 
Enzymatic digestion and purification: 
The whole dialyzed sample was digested overnight with 1µg of trypsin at 37°C. After 
about 16h, the samples were acidified with 1% formic acid (FA) and the proteins 
extracted and purified by solid phase extraction with Empore SPE Extraction Disk (3M). 
The disks (Ø 0.1 cm) were washed with acetonitrile (1 min), conditioned with methanol 
(1 min) and washed with 0.1% FA solution (1 min) before being added to the samples 
and mixed for three hours to allow protein loading on the disk. Following a brief wash of 
the disk in 0.1% FA solution (1 min), the peptide mixtures were then eluted in 100 µL of 
75:25 (v/v) acetonitrile:0.1% FA. All samples were then dried down on speedvac and 
resuspended in 10 µL of 0.1% FA. 
 
Protein analysis by nano flow liquid chromatography tandem mass spectrometry 
(nanoLC-Orbitrap MS/MS) :  
The samples were injected without further dilution; the injection volume was 1 µL. 
Samples were run in duplicates. The peptides were first loaded onto an in-house packed 
Thermo BioBasic C18 precolumn (30 mm x  75 µm i.d.) after which they were separated 
on an in-house packed analytical column (210 mm x 75 µm i.d.) made of the same 
stationary phase, using a Thermo Scientific Dionex Ultimate 3000 UHPLC system with 
the following gradient: 2% B 0-8 min, 55% B 98 min, 90% B 100-103 min, 2% B 104-
120 min, where buffer A was 0.1% FA in H2O and buffer B was 0.1% FA in acetonitrile 
(CAN). The UHPLC was directly coupled to a Thermo Scientific LTQ Velos Dual 
Pressure Linear Ion Trap mass spectrometer which analyzed the peptides in positive 
mode using the following parameters: MS1 60,000 resolution, 100 ms acquisition time, 
1×106 automatic gain control (AGC), MS2 15,000 resolution, 250 ms acquisition time, 
5×105 AGC, top 8, 30 normalized collision energy (NCE) higher-energy collisional 
dissociation (HCD). 
 
Bioinformatics analysis: 
For each sample, the two fractions were combined into one search to create one output 
file. PEAKS 8.5 (Bioinformatics Solutions Inc.) was used to search the RAW data for 
matches against publicly available sequences in imported UniProt (www.uniprot.org) and 
NCBI (https://www.ncbi.nlm.nih.gov/protein) databases. PEAKS PTM was enabled to 
identify unspecific PTMs.  
A search was conducted against a large Uniprot database containing all mammalian and 
bird entries (Uniprot mammalia + aves, imported June 18th, 2018) to identify the source 
of the leather and possible adhesive such as egg white or yolk. Searches were carried out 
using trypsin as enzyme, one allowed non-trypsin cleavage at any end, one missed 
13 
 
cleavage, peptide mass tolerance (PMS) of 10 ppm, fragment mass error tolerance 
(MS/MS) of 0.02 Da, carbamidomethylation as a fixed modification, and deamidated 
(NQ), hydroxylation (RYFPNKD), and oxidation (M) as variable modifications. A 
maximum of five PTMs (Post-Translational Modification) were allowed.  
In addition, NCBI was used to search specifically for collagen sequences as it contains 
additional entries for recently-sequenced species such as Capra hircus (whose collagen 
proteins are incompletely populated in Uniprot). Because NCBI is not curated as Uniprot, 
the database was limited in size to collagen sequences for all mammals as well as fish: 
NCBI Collagen mammalia + fish (last imported January 06th, 2020). Searches were 
carried out using trypsin as enzyme, one allowed non-trypsin cleavage at any end, three 
missed cleavage, peptide mass tolerance (PMS) of 10 ppm, fragment mass error tolerance 
(MS/MS) of 0.02 Da, carbamidomethylation as a fixed modification, and deamidated 
(NQ), hydroxylation (P), and oxidation (M) as variable modifications. A maximum of six 
PTMs (Post-Translational Modification) were allowed.  
 
Results  
Microscopic and SEM/EDX examination 
Metal threads: 
SEM-EDX analysis detected gold, silver or copper on the metal-wrapped threads (Table 
2). Details of the metal-wrapped threads of object no.1 are shown in Figure 4a and the 
SEM image and EDX data are illustrated in Figures 4b & 4c respectively. Among the six 
samples, three of them are solid metal strips wound around a fibrous core, while the 
remaining three are strips made of metal-coated paper wound around a fibrous core. 
 
Table 2. The elemental compositions and characteristics of metal-wrapped threads 
Object Type of metal Area on SEM- Core Substrate 
no. threads hats EDX thread 
results 
1 Gold-wrapped Eyebrows Gold, Not Paper 
threads Silver tested 
2 Metal-wrapped Decorative Copper Cotton No 
threads motifs substrate  
4 Metal-wrapped Eyebrows Copper  Linen No 
threads substrate 
7 Gold-wrapped Decorative Gold, Silk Paper 
threads motifs Silver 
10 Metal-wrapped Decorative Silver Cotton Paper 
threads motifs 
11 Metal-wrapped Lotus at the Silver Silk No 
threads top substrate 
 
14 
 
   
Figure 4. Gold-wrapped threads on the eyebrows of a tiger hat (object no. 1).  (a)  Details. Microscope 
photo ©Conservation Office; (b)  SEM image of the gold-wrapped thread, at a magnification of 140X, 
11.5 mm working distance and 30 kV accelerating voltage ©Conservation Office and; (c)  EDX result of 
the gold layer of the gold-wrapped thread showing the presence of gold , at 30 kV accelerating voltage 
©Conservation Office  
Under microscopy, the surface morphologies of the gold (Figure 5a) or brown strips 
(foils) (Figure 5b) showed the presence of follicles like holes on the grain sides (Figures 
6a & 7a) with fibrous structures on the undersides (Figures 6b & 7b) (Haines 2006; 
Haines and Barlow 1975).  
The cross sections of the metal strips or foils (Figure 8) showed the presence of three 
layers (Figure 8b); a very thin layer of metal is seen on top of a dark brown layer while 
the bulk of the thread is composed of the thick skin base where collagen bundles are 
apparent. The thickness of the samples ranged from 50 µm to 200 µm.  The brown layer 
can possibly be an adhesive layer for binding the metal and base substrate. As shown in 
Figure 9, an adhesive like layer is noticeable on the base substrate of a gold strip (object 
no. 8) where gold is missing.  
SEM-EDX indicated that the major elements on the gold and brown strips or foils were 
gold and silver respectively (Table 3). Sulphur was also detected on the brown strips and 
foils, indicating that the silver was tarnished and accounted for the colour.  The metal 
lamellae are thus composed of gold or silver which are coated on leather-like substrates.  
Calcium and potassium were detected on the undersides of the strips and foils and those 
elements likely result from the processing of leather (Florian 2006).  
Table 3.  SEM results of metal strips or foils samples 
Object Sample Grain side Underside  
no. composition composition 
1 Gold foil Gold, Silver,  Sulphur 
2 Brown foil  Silver Calcium, Sulphur 
3 Brown strip Silver, Not tested 
Sulphur 
4 Gold foil Gold, Silver Sulphur 
15 
 
5 Gold foil Gold, Silver Calcium, Sulphur, Silicon 
6 Gold strip Gold, Silver Calcium, Sulphur, 
Potassium, Silicon 
7 Gold strip Gold Sulphur 
8 Gold strip Gold, Silver Calcium, Sulphur, 
Potassium 
9 Gold strip Gold, Silver Calcium, Silicon, Sulphur 
11 Brown strip Silver, Calcium, Potassium, 
Sulphur Silicon, Sulphur 
 
  
Figure 5. (a) Gold metal lamellae (object no. 6) and; (b) brown metal lamellae (object no. 11). 
Microscope photos ©Conservation Office 
 
 
  
Figure 6. Microscope photos of gold lamellae (object no. 6, samples from ear)  showing the 
presence of follicles (a) on the grain side and;  (b) fibrous structure on the underside  
©Conservation Office 
 
  
16 
 
Figure 7.  SEM images of gold metal lamellae (object no. 6, samples from ear ) (a) grain 
side, at a magnification of 160X, 10 mm working distance and 30 kV accelerating voltage 
©Conservation Office and; (b) underside, at a magnification of 200X, 10 mm working 
distance and 30 kV accelerating voltage ©Conservation Office 
 
 
 
 
  
Figure 8.  Cross section images of gold lamellae (object no. 6, samples from embroidery) 
under (a) SEM, at a magnification of 514X, 10 mm working distance and 30 kV accelerating 
voltage ©Conservation Office and; (b) optical microscope ©Conservation Office 
 
  
 
   
 
 
 
 
 
 
 
 
 Figure 9. Microscope photo showing the presence of an adhesive like layer for 
binding the gold layer to the base substrate of a gold foil (object no. 8, sample S4 for 
 proteomic analysis) ©Conservation Office 
   
Feathers:  
EDX analysis on the waterweeds in the mouths of two goldfish of a hat (object no. 7, 
Figure 1g) showed the presence of sulphur. Further examination indicates the features of 
barbules are consistent with ostrich feathers (Dove and Koch 2011). The barbules are 
uniform with distal prongs, without expanded nodes or pigment patterned (Figure 10).  
Bird feathers are made of proteins commonly known as β-keratins, a type of fibrous 
protein like the α-keratins that compose hair, horn, wool, and nails. Keratins have high 
content of cysteine, an amino acid containing sulphur (Hudon 2005; Staroń et al. 2011).  
The waterweeds of the fish hat are thus deduced to be feathers.    
17 
 
 
  
Figure 10.  SEM image of the feathers representing waterweeds in the mouths of two 
goldfish of a hat (object no. 7) showing the presence of barbules, at a magnification of 1000X, 
10 mm working distance and 30 kV accelerating voltage ©Conservation Office 
 
 
X-radiography  
Two embroidered women figures on a tiger hat (object no. 4) (Figures 11a & 12a): 
The regions decorated only with “leather gold” were darker in intensities than those with 
embroideries, which appeared whiter.   The intensities of the regions with embroidery 
threads or without threads (i.e. paper) were largely the same (Figures 11b & 12b).  There 
was no distinct boundary between the “leather gold” and the embroidered motifs which 
indicated that “leather gold” was a continuous film across the underside of the 
embroideries (with paper).  This shows that the base layer of the embroideries is a gold 
layer on leather substrate, it was covered with a layer of paper and then surface 
embroideries.  The embroidered figures were thus made with layers of gold coated on 
leather, paper and embroideries and sewn onto the surface of a black fabric of the hat.   
 
 
18 
 
b 
 
 
Figure 11. (a) Details of a red woman motif decorated with leather gold on a tiger hat (object no. 
4). photo ©Hong Kong Heritage Museum and; (b) X-ray image of the red woman motif 
©Conservation Office 
 
 
 
Figure 12. (a) Details of a blue woman motif decorated with leather gold on a tiger hat (object 
no. 4), photo ©Hong Kong Heritage Museum and;  (b) X-ray image of the blue woman motif  
©Conservation Office 
 
 
19 
 
A three-dimensional (3D) lotus ornament with a boy figurine on object no. 11 (Figure 
13a): 
 X-ray image (Figure 13b) shows that U-shaped cross wires run inside the linings at the 
crown of the hat to maintain the hat’s shape, a bundle of wires are located inside the 3D 
lotus ornament. The inner wires became visible when the lotus petals were opened up 
gently after humidification.  Those wires are largely used to reinforce the positions of the 
petals though the petals themselves are stiff enough to stand alone.  The boy figurine is 
stood on top of those wire clusters but not fixed to them.  
 
  
Figure 13.  (a) Deformed 3D lotus ornament with a boy figurine featuring at the top of a hat 
(object no. 11), photo ©Hong Kong Heritage Museum ; (b) X-ray image showing the internal 
wiring framework of the lotus ornament with a boy figurine ©Conservation Office  
 
 
Proteomics 
 
A general search was conducted against a large Uniprot database containing all 
mammalian and bird entries to identify the source of the leather and possible adhesive, 
such as egg white or yolk.  
 
Substrate identification: 
The main hits were invariably Type I alpha-1 and alpha-2 collagen chains from Ovis 
aries (domestic sheep). However, because Uniprot had incomplete representation of the 
Capra hircus species compared to NCBI, a complementary search was conducted using a 
smaller NCBI database containing collagen proteins for all bovidae as well as fish 
species. Species markers that discriminate sheep from goat were searched to corroborate 
the database identification. Collagen Type I alpha-1 (collagen alpha-1(I)) and alpha-2 
(collagen alpha-2(I)) are the most abundant collagen chains in skin, and with collagen 
Type III alpha-1 (collagen alpha-1(III)) provide the highest number of peptides. SM 
Tables 1 to 3 show the species markers for all three collagen chains that were used to 
discriminate sheep from goat. They are also compared to cow (Bos taurus), the closest 
domestic relative. Three samples, S3, S5, and S6 were identified as goat, while all others 
were confirmed as sheep (Table 4 and 5). Figure 14 shows the mass spectra of the Capra-
20 
 
specific marker in collagen alpha-2(I), 
GPSGEPGTAGPPGTPGPQGFLGPPGFLGLPGSR (Figure 14a) and its equivalent in 
Ovis sp. GPSGEPGTAGPPGTPGPQGLLGAPGFLGLPGSR (Figure 14b). These 
markers have been characterised before to differentiate sheep from goat in bones 
(Buckley et al.  2010). While none of the leather substrates was identified as bovine, 
traces of Bos taurus collagen were detected in sample S8 with four Bos sp. peptides 
identified from collagen alpha-2(I) (SM Table 2) and one from collagen alpha-1(III) (SM 
Table 3). This is likely part of the adhesive and not the leather itself. 
 
Table 4. List of samples and proteomics identification of the leather substrate 
Sample Object no. Metal Leather identification 
S1 9  Gold Sheep 
S2 9  Gold Sheep 
S3 8  Gold Goat 
S4 8  Gold Sheep 
S5 11 Silver Goat 
S6 11 Silver Goat 
S7 7  Gold Sheep 
S8 2  Silver Sheep 
S9 6  Gold Sheep 
S10 6  Gold Sheep 
S11 6  Gold Sheep 
 
 
Table 5. General results obtained from the NCBI Collagen bovidae + fish (imported 
January 06th, 2020) search on PEAKS PTM giving the number of PSM (Peptide-
Spectrum Matches) and number of peptide sequences identified. Details are given for the 
main three collagen chains identified Col1A1 (collagen alpha-1(I)), Col1A2 (collagen 
alpha-2(I)) and Col3A1 (collagen alpha-1(III)): Score in -10lgP, %=percentage coverage, 
#=number of peptides 
 PSM Peptide Species id Col1A1 Col1A2 Col3A1 sequences Score % # Score % # Score % # 
Ovis aries 181.52 60 92 168.69 62 67 151.26 45 53 
S1 3732 1076 Larimichthys 
crocea 170.96 60 99 163.87 63 76 - - - 
Ovis aries 166.32 58 78 155.09 59 59 137.36 45 50 
S2 5611 1617 Larimichthys 
crocea 178.67 63 192 165.39 69 127 - - - 
Capra hircus 187.34 60 101 168.96 62 94 157.01 46 70 
S3 5506 1526 Larimichthys 
crocea 174.14 61 139 165.49 67 101 - - - 
Ovis aries 184.96 59 99 170.13 58 67 147.3 44 53 
S4 3904 1179 Larimichthys 
crocea 176.2 60 108 173.55 64 93 - - - 
S5 4105 1213 Capra hircus 187.54 60 103 178.18 68 108 166.26 50 76 
21 
 
Larimichthys 
crocea 172.71 58 110 164.46 64 81 - - - 
Capra hircus 191.79 60 104 178.1 65 104 146.46 50 73 
S6 4216 1275 Larimichthys 
crocea 175.35 57 117 164.67 63 85 - - - 
Ovis aries 177.06 58 96 166.62 64 78 150.57 45 50 
S7 5078 1455 Larimichthys 
crocea 179.71 61 140 171 67 110 - - - 
Ovis aries 191.88 63 92 183.92 63 72 158.11 46 60 
S8 3867 1106 Larimichthys crocea 157.05 49 61 149.39 53 53 - - - 
Bos taurus 189.03 60 84 179.59 58 82 139.74 38 51 
Ovis aries 184.11 62 93 172.35 64 71 151.43 48 56 
S9 4472 1214 Larimichthys 
crocea 168.2 60 105 166.12 65 81 - - - 
Ovis aries 164.85 58 79 153.77 58 55 115.52 37 38 
S10 4352 1240 Larimichthys 
crocea 173.13 62 152 164.11 68 114 - - - 
Ovis aries 183.88 59 101 169.08 62 75 152.98 45 54 
S11 3914 1140 Larimichthys 
crocea 167.33 56 104 160.1 62 78 - - - 
 
Adhesive identification: 
The search for the nature of the adhesive did not indicate the presence of non-collagenous 
proteins, such as egg or milk proteins. However, the NCBI collagen database containing 
fish revealed the presence of collagen from the species Larimichthys crocea or large 
yellow croaker, a marine fish from the croaker family (Sciaenidae), native to the 
northwestern Pacific Ocean. The species is important commercially and for this reason 
has had its genome mapped (Wu et al. 2014). In spite of the limited representation of fish 
collagen in public database, L. crocea collagen is identified in all samples with high 
number of peptides (Table 5, SM Figures 1 and 3) and similar protein coverage than the 
substrate collagen (e.g. 62% on average for fish collagen compared to 63% on average 
for the substrate for collagen alpha-2 (I)), giving high confidence that the collagen is from 
a species from the Larimichthys genus. The Larimichthys genus is composed of at least 
three species: L. crocea (large yellow croaker) (Cui et al. 2009), L. polyactis (small 
yellow croaker found in Yellow and East China seas) (Cheng et al.  2012), and L. 
pamoides (southern yellow croaker found in Australia and Papua New Guinea) (Munro 
1964). A fourth species, L. terengganui, has recently been discovered in Peninsular 
Malaysia (Seah et al. 2015). Of all the Larimichthys, only L. crocea has been completely 
sequenced, while only mitochondrial DNA is available for L. polyactis. Possible peptides 
specific to L. crocea are shown in SM Tables 4 and 5, and SM Figures 2 and 4.  
 
The identification of Larimichthys crocea is rather similar for all samples, with the 
exception of sample S8, where the protein percentage coverage is the lowest. As noted 
earlier, Bos taurus was also identified in this sample, albeit with less markers than Ovis 
aries. This indicates that the adhesive in S8 is likely a mixture of fish and bovine glue, 
and the lower coverage of the fish glue perhaps indicates a lower supply, which was 
compensated by the addition of hide glue from bovine.     
 
22 
 
 
a) y ions
b ions
 
23 
 
b) y ions
b ions
 
 
Figure 14. Mass spectra showing the b and y ions series corresponding respectively to 
peptide fragments from the amino terminus (extending from the left) and the carboxyl 
terminus (extending from the right) of (a) Peptide 
GPSGEP*GTAGPP*GTP*GPQGFLGPP*GFLGLP*GSR with 5 hydroxyprolines (P*) 
in sample S5: Molecular mass of the peptide is M=3092.4839, identified at a mass-to-
charge ratio m/z=1031.8416 and PEAKs score of 46.15 (top) and; (b) Peptide 
GPSGEP*GTAGPP*GTP*GPQGLLGAP*GFLGLP*GSR with 5 hydroxyprolines (P*) 
in sample S10: Molecular mass of the peptide is M=3032.4839, identified at a mass-to-
charge ratio m/z=1011.8407 and PEAKs score of 49.00 (bottom), obtained by nanoLC-
Orbitrap MS/MS and analyzed in PEAKS 8.5 ©Smithsonian’s Museum Conservation 
Institute 
 
Discussion 
As a result of the proteomic analysis, it was confirmed that “sheep and goat skins” are 
used as the base substrates of the “membrane threads” of the traditional Chinese 
children’s hats.  “Sheep leather gold” is a commonly known membrane thread in China.   
Both sheep and goat skins were found on a fish-shaped hat (object no. 8).  This indicates 
that the raw materials of “sheep leather gold” could come from sheep or goat in the early 
20th century.   Moreover, fish collagen was also found in all samples and it was 
enlightening that the species of fish, yellow croaker, was matched with high confidence.  
The fish collagen is likely used in the form of fish glue to bind the metal onto the leather, 
24 
 
as samples have noticeable adhesive layers beneath the metal layers when viewed under 
microscope (Figure 9). However, it cannot be ruled out that the glue was also used for 
fixing the “membrane threads” to the hats.  Some of the metal strips are indeed adhered to 
the hats with glues rather than stitching. Nevertheless, the two findings correlate well 
with the ancient materials that were used for producing “leather gold” as well as its 
application as documented in the ancient encyclopaedia.  
The use of “leather gold” dating to the 17th century was also reported in other texts, 
ranging from literary to historical and medical. It was described as goods and available in 
the trade market at that time.  In a well-known Chinese novel “The Plum in the Golden 
Vase” (Jin Ping Mei), the hems of woman’s costumes were decorated with “leather gold”. 
More curiously, it could be used to cure fracture or injuries in Mongolia as cited in the 
ancient medical book “Supplement to Compendium of Materia Medica” (Temule 2011). 
 Leather gold embroidery is also known as “shadow embroidery”.  Leather gold is affixed 
as the backing of a motif, whereas suitable embroidery stitches would then be used to 
form the motifs.  The sparkling effect of gold can be illustrated among the embroideries.  
This technique is said to be similar to an ancient embroidery technique which was 
described in a historical text “History of Song” (AD 960–1279, the Song dynasty).  The 
embroidery work was combined with gold gilding to show the glittering of gold on the 
embroideries (Meng 2007).  
Yellow croaker was identified which is likely one or the only component of the adhesive 
to bind the gold or silver strips (foils) onto the leather.  There has been a long history of 
using fish glue in China since 475 BC (Pan et al. 2018). Dried swim bladder from the 
yellow croaker have been used to make fish glue (Lin 1939). It has been widely used in 
joining wood, especially for mortise and tenon joints, since ancient China (Zhang 2011).  
The use of fish glue was considered to be stronger and reversible in nature (Schellmann 
2009). Tracing back to the record in the 17th century, fish glue made from the bladders 
and intestines of fish had been used to make bows,  and Seiaena schlegeli (a now 
replaced scientific name designating a croaker fish) was used in Chekiang (nowadays 
Zhejiang Province)  (Sung 1966).   Various animal glues, including fish glue, were also 
implemented as binders in paint media during the Tang dynasty (AD 618-906) 
(Petukhova 2000).  
Based on a Chinese online article written by the curator of Jincheng Museum about the 
making of “leather gold” in Jincheng of Shanxi in the olden days (Qing dynasty, AD 
1644-1911), fish swim bladders, cape jasmine (Gardenia jasminoides) and Chinese gall 
(Rhus chinensis)  were used.  The latter two materials could possibly be used as 
antibacterial and tanning agents respectively (Guo 2011). Those literature findings on fish 
glue and leather gold reiterate the consistency of using traditional materials from the 
ancient time for the production of the children hats in the early 20th century. 
 
25 
 
Conclusion 
The use of analytical tools such as microscopy, SEM-EDX, X-radiography and 
proteomics techniques provided insights about the decorative materials and fabrication 
techniques of these traditional children hats. In addition to textile fabrics like cotton and 
silk, the hats are composed of a variety of materials, in particular the metal-coated strips 
or foils made of organic materials. The strips or foils are made of gold or silver likely 
adhere onto sheep or goat leathers with fish glue from the croaker family.  The leather 
types and glue were identified via proteomics, a powerful analytical technique for the 
species identification of skin and other animal tissues in minute amounts, notwithstanding 
still a destructive technique.  The identification of adhesive is always challenging as the 
adhesive may not be easily isolated for analysis depending on its quantity, purity or 
structural condition.         
Overall, the study provides valuable material information from historical and preservation 
perspectives.  The composite nature of the children hats requires a stable environment to 
retard their deterioration, as well as good physical support for maintaining the shapes in 
storage.  The recognition of the material types of the “gold” strips at the treatment stage 
cautioned the conservators to humidify the hats in a controlled manner for reshaping.  
The materials for fabricating the “leather gold” are largely in line with those used in 
ancient times, suggesting the making of “leather gold” could have also been inherited 
from the traditional ways.  This reflects that the craftsmanship of children hats has passed 
down through families, from generation to generation, within the traditional cultural 
context. 
 
Acknowledgment 
Thanks are due to staff of Conservation Office, Ronnie Kam, Iris Lo ( former staff) and 
Tammy Cheung (former conservation assistant) for conducting SEM-EDX analysis and 
Jonathan Tse for X- radiography.  Special thanks goes to Evita Yeung for her support of 
the study.   The authors would also like to express special thanks to Mary Ballard, MCI 
who proposed the use of proteomic analysis for identifying the leather species. The first-
of-its-kind collaboration between Conservation Office, LCSD, HKSAR with MCI was 
thus established for the proteomics. The proteomics analyses were carried at MCI’s 
Proteomics and Molecular Mass Spectrometry Laboratory with the help of Drs. Timothy 
Cleland and Asher Newsome, and supported by Smithsonian’s Museum Conservation 
Institute Federal and Trust Funds and the Andrew W. Mellon Foundation – Directorship 
Endowment (Grant # G-40700657). 
References 
British Museum a. 2020.  https://www.britishmuseum.org/collection/object/A_MAS-908-
b, https://www.britishmuseum.org/collection/object/A_MAS-908-a [accessed 26 August 
2020] 
26 
 
 
British Museum b. 2020.  https://www.britishmuseum.org/collection/object/A_MAS-911 
[accessed 26 August 2020] 
 
Brooks, M.M., and S.A. O'Connor. 2007. X-Radiography of Textiles, Dress and Related 
Objects. Edited by M.M Brooks, and S.A. O'Connor. Oxford, UK. Elsevier : Butterworth-
Heinemann. 
Buckley, M., S. Whitcher Kansa, S. Howard, S. Campbell,  J. Thomas-Oates, and M. 
Collins. 2010. “Distinguishing Between Archaeological Sheep and Goat Bones Using a 
Single Collagen Peptide.”  Journal of Archaeological Science 37(1): 13-20. 
Cao, Y. 2010. “Tiger-Head Cap of Art and the Aesthetic.”   Art and Design 04: 286-288.  
[in Chinese] 
Cheng Y., R. Wang, Y. Sun, and T. Xu. 2012. “The Complete Mitochondrial Genome of 
the Small Yellow Croaker and Partitioned Bayesian Analysis of Sciaenidae Fish 
Phylogeny.” Genetics and Molecular Biology: 35(1) 191-199. 
 
Cui Z., Y. Liu, C.P. Li, F. You, and K.H. Chu, 2009. “The Complete Mitochondrial 
Genome of the Large Yellow Croaker, Larimichthys crocea (Perciformes, Sciaenidae): 
Unusual Features of its Control Region and the Phylogenetic Position of the Sciaenidae.” 
Gene, 432(1-2):33-43. 
 
Dallongeville, S., M. Koperska, N. Garnier, G.eve Reille-Taillefert, C. Rolando, 
and C. Tokarski. 2011.  “Identification of Animal Glue Species in Artworks Using 
Proteomics:  Application to a 18th Century Gilt Sample.” Analytical Chemistry 83: 9431-
9437. 
 
Dallongeville, S., M. Richter, S. Schäfer, M. Kühlenthal, N.  Garnier, C. Rolando, and C.  
Tokarski. 2013. “Proteomics Applied to the Authentication of Fish Glue: Application to a 
17th Century Artwork Sample.” The Analyst 138: 5357-5364. 
 
Dallongeville, S., N.Garnier, C.Rolando, and C.Tokarski.  2015.  “Proteins in Art, 
Archaeology, and Paleontology: From Detection to Identification.”   Chemical Review 
116 (1): 2–79. 
Dang, X.J., J.L. Guo, K. Bai, and J.C. Yang. 2014. “The Study of Twining Gold Thread 
From the Shanpula Tomb in Luopu County, Xinjiang Uyghur Autonomous Region of 
China.” Science of Conservation and Archaeology 26(3): 13-18. [in Chinese] 
Dove, C.J., and S. L. Koch. 2011. “Microscopy of Feathers: A Practical Guide for 
Forensic Feather Identification.”  The Microscope 59(2): 51-71. 
27 
 
Florian, M-L. E. 2006. “The Mechanism of Deterioration in Leather.” In Conservation of 
Leather and Related Materials, edited by K. Marion and R. Thomson, 40. Oxford: 
Butterworth-Heinemann. 
Garrett, V. M.  2008. “Clothing for Children.” In Chinese Dress: From the Qing Dynasty 
to the Present.”  180-195. Tuttle Publishing. 
Guo, L. 2011.  The Hand Crafting Industrial of Jincheng in the Ming and Qing Dynasty:  
The Rise and Decline of the Leather Gold Industrial.  Accessed 31 August, 2017. [in 
Chinese]  
Hacke, A.M., C.M. Carr, and A. Brown. 2004. “Characterisation of Metal Threads in 
Renaissance Tapestries.” In Metal 04, Proceedings of the International Conference on 
Metals Conservation 4-8 October”, edited by J. Ashton, and D. Hallam, 415-425. 
Canberra, Australia. 
Hackett, J. 2011. “X-radiography as a Tool to Examine the Making and Remaking of 
Historic Quilts.”  Victoria and Albert Museum Online Journal 3.  
Haines, B.M. and J.R. Barlow. 1975. “The Anatomy of Leather.”  Journal of Materials 
Science 10(3): 525-538. 
Haines, B.M. 2006. “The Fibre Structure of Leather.” In Conservation of Leather and 
Related Materials, edited by K. Marion, and R. Thomson, 11-21. Oxford: Butterworth-
Heinemann. 
Hložek, M., T. Trojek, R. Prokeš, and V. Linhart. 2019. “Mediaeval Metal Threads and 
Their Identification Using Micro-XRF Scanning, Confocal XRF, and X-ray Micro-
radiography”. Radiation Physics and Chemistry 155: 299-303. 
 
Hong Kong Heritage Museum (HKHM). 2015.  Wearable Blessings: Traditional Chinese 
Children’s Clothing Exhibition.   
Hu, X.R., and W.D. Yu.  2016.  “Technology and Category of China Ancient Silk 
Decorated in Gold and Historical Inheritance.” Journal of Textile Research   37(8): 65-
71. [in Chinese] 
Hudon, J. 2005. “Considerations in the Conservation of Feathers and Hairs, Particularly 
their Pigments.”  In Fur Trade Legacy. The Preservation of Organic Materials, edited by 
M. Brunn, and J. A. Burns, 127-147. Canadian Association for Conservation of Cultural 
Property, Ottawa, Canada. 
Indicator, N., J. Koestler, M. Wypyski, and A.E. Wardwell. 1989. “Metal Threads Made 
of Proteinaceous Substrates Examined by Scanning Electron Microscope – Energy 
Dispersive X-ray Spectrometry”. Studies in Conservation 34: 171-182.  
 
Járó, M. 1990. “Gold Embroidery and Fabrics in Europe XI-XIV Centuries.” Gold 
Bulletin 23 (2): 40-57. 
28 
 
Járó, M, and A. Tóth. 1991. “Scientific Identification of European Metal Thread 
Manufacturing Techniques of the 17-19th Centuries.”  Endeavour 15 (4): 175-84. 
Járó, M., T. Gál., and A. Tóth. 2000. “The Characterization and Deterioration of Modern 
Metallic Threads”. Studies in Conservation 45 (2):  95-105. 
Karatzani, A., Z.Y. Lu, and R. Thilo. 2009. “The Metal Threads from the Silk Garments 
of Famen Temple.”  Restaurierung und Archäologie 2: 99-109. 
Karatzani, A.  2012.  “Metal Threads: the Historical Development.” Accessed 
31December, 2019. 
https://conferences.saxo.ku.dk/traditionaltextilecraft/keynote_speakers/presentations/Ann
a_Karatzani.pdf 
Kohara, N., Y.  Sasa, K. Sakurai, and M. Uda. 1998. “A Note on the Characterization of 
Metal threads in Historic Textiles Handed Down by the Ainu People.”  Studies in 
Conservation 43 (2): 109-113. 
Kumazawa,Y., Y. Taga, K. Iwai, and Y.I. Koyama.  2016. “A Rapid and Simple LC-MS 
Method Using Collagen Marker Peptides for Identification of the Animal Source of 
Leather.”  Journal of Agricultural and Food Chemistry 64: 6051-6057. 
 
Leo, G., L. Cartechini, P. Pucci, A. Sgamellotti, G. Marino, and L. Birolo.  2009. 
“Proteomic Strategies for the Identification of Proteinaceous Binders in Paintings.” 
Analytical and Bioanalytical Chemistry 395(7): 2269–2280. 
 
Leung, Y.S. 2010. Hidden Meanings in Chinese Children’s Clothing and Accessories. 
Guangzhou Publishing House. [in Chinese]  
Lin, P. L. 2007.  Symbolism in Chinese Children’s Hats and Baby Carriers: Folklore, 
Bonding, and Mother’s Affectionate Embrace.  Accessed 1 November, 2017. 
https://www.uindy.edu/asian-programs/files/outline_symbolism_in_chinese_children.pdf 
Lin, S. Y. 1939. “Fish Air-bladders of Commercial Value in China.” The Hong Kong 
Naturalist  9: 108-118. 
Lu, J., L. Niu, and R.R. Cui. 2016. “Modern Decorative Elements of Han Folk Children 
Bonnets.” Journal of Textile Research 37(4): 119-123. [in Chinese] 
Lu, Z.Y.  2015. “The Study of the Gold Decorative Techniques on Silk Textiles of Tang 
Dynasty Uunearthed from the Famen Temple.” Archaeology and Cultural Relics 6: 110-
120. [in Chinese] 
Lu, Z.Y., R. Hui, and S.M. Han.  2017. “Comparative Morphological, Compositional and 
Technical Research on Three Groups of Ming and Qing Dynasty Metallic Threads.” 
Science of Conservation and Archaeology 29 (4): 36-44. [in Chinese] 
29 
 
Luo, W.G., Y. Si, H.M. Wang, Y. Qin, F.C Huang, and C.S Wang. 2011. “Leather 
Material Found on a 6th B.C. Chinese Bronze Sword: A Technical Study.” 
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 79 (5): 1630-
1633. 
Luo, X.Y. 2018. “The Format and Decorative Features of Animal Shaped Children’s Hats 
of the Late Qing Dynasty.” Anhui Literature 2: 63-64. [in Chinese] 
Ma, Z., J. Yan, X. Zhao, L. Wang, and L. Yang. 2017. “Multianalytical Study of the 
Suspected Binding Medium Residues of Wall Paintings Excavated in Tang Tomb, 
China.” Journal of Cultural Heritage 24: 171–174. 
 
Mao, H.W.  2020. “Comparative Study of Nasij and Jin Duan zi of Mongol Period.” 
Asian Social Science 16 (5): 13-18. 
 
Matar, R., and M. Merheb. 2016. “Paleogeneticist View of Leather: The Role of 
Mitochondrial DNA to Uncover the Mysteries of Fake leather and its Products.” Biosci. 
Biotech. Res. Comm. 9(1): 6-14. 
Meng, H. 2007. “Something on Gold Embroidery.” Space 10: 38-39.  [in Chinese]  
Merheb,  M., S. Vaiedelich , T. Maniguet, and C. Hänni. 2014. “Molecular Species 
Identification in Processed Animal Hides for Biodiversity Protection.”   
International Journal of Advances in Chemical Engineering and Biological Sciences 
1(1): 55-57. 
Munro, I.S.R. 1964. "Additions to the Fish Fauna of New Guinea." Papua and New 
Guinea Agricultural Journal 16(4): 141-186. 
 
Muros, V., S.K.T.S. Wärmländer,  D.A. Scott, and J.M.  Theile. 2007. “Characterization 
of 17th-19th Century Metal Threads from the Colonial Andes.” Journal of the American 
Institute for Conservation 46(3): 229-244. 
 
Nord, A. G., and K. Tronner. 2000. “A Note on the Analysis of Gilded Metal Embroidery 
Threads.” Studies in Conservation 45(4): 274-9. 
Oraltay, R.G., and R. Karadag. 2020. “Surface Investigation of Metal Threads and Solid 
Metals of Ottoman Textiles in the Topkapi Palace Museum.” Studies in Conservation 65 
(1):  59-64. 
Pan, L., P. Li, Y. Tao, and H.R. Guo. 2018. “Antimicrobial Agent Effectiveness in Fish 
Glue Prepared by Heat Treatment and Enzymatic Hydrolysis of Swim Bladders.” 
BioResources 13(2): 3275 -3283.  
Petukhova, T. 2000.   “A History of Fish Glue as an Artist’s Material: Applications in 
Paper and Parchment Artefacts.”   The Book and Paper Group Annual 19: 111-114. 
Assessed 31 December, 2019.  
https://cool.conservation-us.org/coolaic/sg/bpg/annual/v19/bp19-29.html 
30 
 
Popowich A.K., T.P. Cleland, and C. Solazzo. 2018. “Characterization of Membrane 
Metal Threads by Proteomics and Analysis of a 14th c. Thread from An Italian Textile.” 
Journal of Cultural Heritage 33: 10-17. 
Rezić, I., L. Ćurković, and M. Ujević. 2010. “Simple Methods for Characterization of 
Metals in Historical Textile Threads.”  Talanta 82: 237-244. 
Schellmann, N. 2009.  “Animal Glues – Their Adhesive Properties, Longevity and 
Suggested Use for Repairing Taxidermy Specimens.”  Natural Sciences Collections 
Association News 16: 36-40. 
Scibè C., C. Solazzo, I. Tosini, T. Lam, E. Vicenzi, M.J. González López. 2020. “Gilt 
Leather Threads in 11th-15th Century Textiles”. In Proceedings of the 11th Interim 
Meeting of The Leather & Related Materials Working Group, 6-7 June 2019, in Press. 
Paris, France. 
Seah Y.G., N., Hanafi, A.G. Mazlan, and N.L. Chao 2015. “A New Species of 
Larimichthys from Terengganu, East Coast of Peninsular Malaysia (Perciformes: 
Sciaenidae).” Zootaxa 3956(2): 271-280. 
 
Shao, Y. 2005. “Chemical Analysis of Leather”. In Chemical Testing of Textiles, edited 
by Q.G.  Fan.  47-71. Woodhead publishing. 
Solazzo, C., M. Wadsley, J. M. Dyer, S. Clerens, M. J. Collins, and J.  Plowman.  2013. 
“Characterisation of Novel α-Keratin Peptide Markers for Species Identification in 
Keratinous Tissues Using Mass Spectrometry.” Rapid Communications in Mass 
Spectrometry 27(23): 2685-2698. 
Solazzo, C., B. Courel, J. Connan, B.E. van Dongen, H.  Barden, K.  Penkman, S. Taylor, 
B. Demarchi, P.  Adam, P. Schaeffe, A. Nissenbaum, O. Bar-Yosef, and M. Buckley. 
2016. “Identification of the Earliest Collagen- and Plant-based Coatings from Neolithic 
Artefacts (Nahal Hemar Cave, Israel).” Scientic Reports 6: 1-11. 
Solazzo, C. 2019. “Characterizing Historical Textiles and Clothing with Proteomics.”  
Conservar Património 31: 97-114. 
 
Solazzo, C., C. Scibè, and K. Eng-Wilmot. 2019. “Proteomics Characterization of 
"Organic" Metal threads - First Results and Future Directions.”= Research and Technical 
Studies Specialty Group Postprints from the 47th Annual Meeting of the American 
Institute for Conservation in New England, May 13 – 18, 2019, edited by M. McGath, 
Vol 7: 78-82. Washington, DC: The Research and Technical Studies Group (RATS) and 
the American Institute for Conservation (AIC). 
 
Staroń, P., M. Banach, and Z. Kowalski. 2011. “Keratin-origins, Properties, Application.” 
CHEMIK 65(10): 1019-1026. 
31 
 
Sun, N. N., and J. Q. Zhang. 2008. “Cultural Characteristics of the Tiger Pattern Dress.” 
Journal of Textile Research 29:4 108-111. [in Chinese] 
Sung, M.Y.M. 1979.  “Chinese Language and Culture: A Study of Homonyms, Lucky 
Words and Taboos.” Journal of Chinese Linguistics 7 (1): 15-28. 
Sung, Y.H., S. C. Sun (translator), and E. Z. Sun (translator). 1966.  Chinese Technology 
in the Seventeenth Century: Tien Kung Kai Wu. 69-72, 237, 262.  University Park: 
Pennsylvanian State University.  
Szeto, N. Y.Y., and V. M. Garrett. 1990. Children of the Gods Dress and Symbolism in 
China. Hong Kong: Hong Kong Museum of History.   
Temule, T.  2011.  “Leather Gold Part 1 and 2.” Journal of Chinese Literature and 
History. 2: 356,378. [in Chinese] 
Thomson, R. 2006. “Testing Leathers and Related Materials.” In Conservation of Leather 
and Related Materials, edited by K. Marion, and R. Thomson, 59. Oxford: Butterworth-
Heinemann.   
Tokarski, C., E. Martin, C. Rolando, and C. Cren-Oliv. 2006. “Identification of Proteins 
in Renaissance Paintings by Proteomics.” Analytical Chemistry 78: 1494–1502. 
 
Vinciguerra, R., A.De Chiaro, P. Pucci , G. Marino, and L. Birolo. 2016.  “Proteomic 
Strategies for Cultural Heritage: From Bones to Paintings.” Microchemical Journal 126: 
341-348. 
 
Wang, L., and F. Zhao. 2012.  “Exploring the Variation and Development of Embroidery 
in Tang Dynasty Based on the Embroidery Discovered in Dunhuang.”  Silk 49: 60- 65.  
[in Chinese] 
 
Weiszburg, T., K. Gherdán, K.  Ratter, N. Zajzon, Z.  Bendő, G. Radnóczi, A. Takács, 
T.  Váczi, G. Varga, and G. Szakmány. 2017.  “Medieval Gilding Technology of 
Historical Metal Threads Revealed by Electron Optical and Micro-Raman Spectroscopic 
Study of Focused Ion Beam-Milled Cross Sections.” Analytical Chemistry 89 (20): 
10753-10760.  
 
Wu, C.W., D. Zhang, M. Y. Kan, Z. M. Lv, A.Y. Zhu, Y.Q. Su, D.Z. Zhou, J.S. Zhang, 
Z. Zhang, M.Y. Xu, L.H. Jiang, B.Y. Guo, T. Wang, C.F. Chi, Y. Mao, J.J. Zhou, X.X. 
Yu, H.L. Wang, X.L. Weng, J. G. Jin, J. Y. Ye, L. He, and Y. Liu. 2014. “The Draft 
Genome of the Large Yellow Croaker Reveals Well-developed Innate Immunity.” Nature 
Communications 5: 5227. 
Yang, J.C., J. Zhang, and J. Jiang. 2013. “The Making Technique of the Gold-plated 
Thread Unearthed from the Crypt of the Famen Temple.” Archaeology 2:99-104. [in 
Chinese] 
 
32 
 
Zhang, X. M.  2011.  Chinese Furniture. 46-47. Cambridge University Press.  
Zhao, F. 2005.  A History of Chinese Silk Art. 18,72. Beijing: Wenwu Press. [in Chinese] 
Zhao, F. 2012.  “Gold Nasij -Weaving Exchange Between China and Persia During the 
Mongol and Yuan Dynasty.” In Chinese Silk and the Silk Road, 277-304.   City 
University of Hong Kong Press. [in Chinese] 
Zhao, F.  2015. “Weaving Technology.” In A History of Chinese Science and Technology 
Volume 2, edited by Y.X. Lu, 447.   Shanghai: Shanghai Jiao Tong University Press and 
Springer-Verlag Berlin Heidelberg.    
Zhu Z., P. Ta, J.C. Yang, H. Ge, and L. Liu. 2019. “Mass Spectrometric Identification of 
Adhesive Utilized in a Tian-tsui Tiara of the mid-Qing Dynasty (1776–1839 CE) in the 
Collection of the Tang Clan Folk Museum.”  Studies in Conservation 64(4): 187-192. 
 
 
33