The Cast Iron Lion of Cangzhou

Donald B. Wagner

Needham Research Institute newsletter,
no. 10, June 1991, pp. 2–3.

Web version 1999, stylistic revisions 2022, 2024.

In this Web version I have put back some material that was excised then because of lack of space.

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Monumental iron castings – pagodas, bells, statues – were produced in China as early as the Tang period and continued to be produced into modern times, but it seems to have been in the early Song that they were especially popular. Said to be the largest iron casting on record in the world is the famous Cangzhou Lion, cast in A.D. 953. It is 5.4 metres high, 5.3 metres long, and 3 metres broad; its weight is estimated at 50 tonnes. Figure 1 is a photograph of it taken by the American mining engineer Thomas T. Read about 1910. The other photographs were taken by me on a visit to Cangzhou in 1987.

Art historians believe that the lion was originally inside a Buddhist temple, now long gone, and that a bronze statue of the bodhisattva Mañjuśrī rode in the lotus flower on the lion's back. The bodhisattva was at some time removed for the value of its bronze; this may have happened as early as the reign of Shizong (r. 954-958), Emperor of the minor dynasty of Later Zhou, in his campaign against Buddhism.

By 1603 the tail was missing. The lion toppled over in a storm in 1803, with the result that the snout and belly were damaged. In 1886 the Department Magistrate Gong Yu sent masons to prop it up with bricks and stones. When Read saw it the casting was broken into four parts: the body, the lotus seat, the head, and the lower jaw, which lay on the ground and cannot be seen in the photograph. It has not been seen since. In 1984 the lion was professionally restored and placed on a reinforced concrete pedestal two metres high (Figures 2–4).


The chemical analysis and metallographic examination of a sample taken by Read was summarily reported in 1938, and recently a team of Chinese archaeologists and metallurgists has published a more thorough study of the way in which the lion was cast. When I visited Cangzhou a local official helped me to climb up on the pedestal to examine the casting more closely; within a very few minutes, however, I was angrily ordered down by another official. The following description is therefore based almost solely on the published accounts and not to any great extent on personal examination.

The rectangular grid of casting seams on the surface of the Lion has led some art historians to believe that it was somehow pieced together from separately cast iron plates with dimensions about 25×45 cm. Actually the entire statue was cast, essentially in one piece, in one mould, using the traditional Chinese "piece-moulding" method. The seams show the joins between the individual pieces of the mould.

The piece-moulding method was used in bronze-casting in China as early as the Shang period, and to some extent it continues to be used in bronze- and iron-casting today. First an exact full-scale model is made of the intended casting. This is of clay, suitably reinforced with wooden or metal supports. When the model is dry it is plastered to a considerable thickness with wet clay; when this clay has dried to a leather-hard consistency it is cut into blocks which are carefully removed, retaining the impression of the model. These separate blocks are dried and perhaps fired for greater durability. In order to form the mould-core the model is then carefully scraped down, removing a thickness of clay corresponding to the intended thickness of the casting. The previously formed blocks are reassembled around the core to form the outer mould. This is reinforced with a wooden framework and buttressed with earth. One or more casting inlets are made, and the molten metal is poured into the mould thus formed. When the mould is removed the casting shows obvious seams at the joins between the individual blocks of the outer mould. In bronze castings these are usually removed by grinding and polishing, but in iron castings this is normally not possible and the seams are simply left as they are.

When, in the piece-moulding of the Cangzhou Lion, the outer mould was built up around the core, numerous round-headed wrought-iron spikes were driven into the sides of the core to establish the correct spacing. On the back, cast-iron blocks (called ‘chaplets’ by foundrymen) were used for the same purpose. When the iron was poured into the mould the spikes and chaplets were incorporated into the casting; their traces can be found by close examination of the outer and inner surfaces.

Examination also shows that a reinforcing framework of wrought iron was cast into the neck and back of the lion (Figure 3). Presumably this was done because cast iron alone would not have been strong enough to carry the five-ton lotus and the bronze bodhisattva.

Horizontal lines of what foundrymen call ‘cold shuts’ are found at regular intervals from about middle height upward. A cold shut is a fault in a casting caused by premature cooling in the mould before the mould is filled. These indicate that the lion was cast in several stages. First the outer mould was built up to about half of the total height and iron was cast into this (Figure 5). Hereafter the building of the rest of the mould and the casting into it proceeded upward in alternating stages. Cold shuts severely weaken a casting. The founders were aware of this problem, and dealt with it by inserting pieces of wrought iron into the solidifying iron surface, after each partial pour, to act as pegs holding the separately cast levels of the casting together. The operation was not entirely successful, and some of the pegs were displaced by the hot running metal from their intended positions. Some have since fallen out of the casting, leaving characteristic cavities or even holes behind.

Two chemical analyses are available for samples from the iron lion, one by Read and one by the Chinese team. These show, respectively:

Carbon %

Silicon %
Phosphorous %
Manganese %
Sulphur %
3.96
0.09
0.231
0
0.022
4.3
0.04
0.087
0.02
0.022

(Neither report indicates what part of the casting the sample for chemical analysis was taken from.) The difference between the two analyses, especially with respect to phosphorous (P), is another clear indication that the casting was not the result of a single uninterrupted pour. The very low silicon content is typical of pre-modern Chinese cast iron: modern ironfounders would prefer an alloy with about 2 per cent silicon for such a casting. Metallographic examination of Read's sample indicates that it is white cast iron, with a streak of mottling. The Chinese team took three metallographic samples of the cast iron fabric of the casting, one from the right side of the lion's back and two from the lotus; two of these were pearlitic grey cast iron and one, from the upper rim of the lotus, was white cast iron with a small amount of mottling.


In the Qing period a local poet, Ji Ruiqi, wrote:

Thinking of ancient flourishing glory,
I sigh for the changes of our times
But the Iron Lion still stands,
While halls and palaces have turned to thorns and brambles.
The Iron Lion is still Cangzhou's greatest glory, and even the local beer (which is excellent) is called Cangzhou Lion Beer.

Literature consulted:

Wanli Cangzhou zhi, 1603, 1: 9b-10a, 12b, 8: 12a, 19a-b, 21b, 22b, 43a-45a.

Cang xian zhi, Tianjin 1933, frontispiece, 13: 1b-2a, text appendix 1: 26b-27a, 2: 1a-2a, 17a.

T. T. Read, The iron age, 30 April 1936, 18-20; Mining and metallurgy, 1937, 18: 383.

M. L. Pinel, T. T. Read, & T. A. Wright, American Institute of Mining and Metallurgical Engineers, Technical publication no. 882, 1938.

Joseph Needham, The development of iron and steel technology in China, London 1958, pp. 20-21, 72, plate 24.

Luo Zhewen, Wenwu 1963.2: 38 + back cover.

Wu Kunyi, Li Jinghua, & Wang Minzhi, Wenwu 1984.6: 81-85.

Wang Minzhi, Wenwu tiandi, 1985.2: 37.

Zhongguo meishu quanji: Diaosu bian, 5: Wudai Song diaosu, Beijing 1988, p. 19.