THE BRITISH LIBRARY'S SADO SCROLLS

HAMISH TODD

AMONG the nearly eleven hundred works acquired by the British Museum from the collection of Philipp Franz von Siebold^ in 1868 were three hand-painted scrolls depicting mining activities on the Japanese island of Sado. The scrolls belong to a genre of manuscripts known as Kinzan emaki #|IJ^# (Illustrated scrolls of mines)^ or Kwginzan emaki 4?IIUJ^# (Illustrated scrolls of gold and mines), which provide a detailed representation of both the underground mine workings and the various processes and administrative operations that went on above ground. Sado, an island measuring 857 square kilometres, lies thirty-five kilometres off the north-west coast of Honshu. For over eleven centuries from 702 it was a province in its own right but on the establishment of the prefectural system of local administration in 1871 it became part of . Today it is well known as the home of the Kodo drummers but in the past its claims to fame were as a place of exile for disgraced notables^ and, more relevant to this article, as a source of gold and other precious . The earliest reference to gold from Sado is to be found in the Konjaku monogatarishu ^^^M^ (Collection of tales now past)'* which dates from the early twelfth century. One of the stories describes how the Governor of Noto, having heard of the abundance of gold on Sado, sends -workers to the island to bring some back. This story is generally thought to refer to gold-panning rather than to actual mining and for many centuries it was this alluvial gold, with some silver, dug from the of the seashore that constituted Sado's riches. During the latter half of the sixteenth century the exploitation of auriferous containing small amounts of gold and silver, mined from the hard rock in underground workings, allowed the island's full potential as a source of precious metals to be developed. Following the arrival in Japan of the Portuguese missionaries and traders, the increased need for gold and silver for trade acted as an incentive for the exploitation of Sado's resources. A merchant called Toyama Mouemon ^UJ/^^f^P^ opened the first mine at Tsurushi in 1542. Then in 1589 the warlord Toyotomi Hideyoshi ®g:^^ (1537-98) ordered the samurai Uesugi Kagekatsu ±.Vi>f^B (1555-1623) to take troops to the island to seize control of the mine. Five years later, anxious to finance his campaign against Korea, Hideyoshi sent mining experts to the island to speed up production. Following the death of Hideyoshi and the rise to power of Tokugawa Ieyasu 130 (1542-1616) the new Tokugawa government also began to take an interest in Sado and its riches. In 1600 the island was placed under the direct control of the Shogunate with an official called the Sado datkan '^miXt: (Deputy for Sado) and later the Sado bugyo \^m.^n (Commissioner for Sado) in charge of operations. Initially the focus of activity was the Tsurushi mine but during the tenure of the first Sado datkan, Tanaka Seiroku 15 4^^7^ , came the discovery of the Aikawa mine which was to prove the most important and profitable of all those on the island. To encourage the development of mining operations the Shogunate introduced a system called unjo nymatsusei '^}LMIV\ (levy bidding) whereby prospectors were given freedom to search for deposits of gold and silver. If they were successful in finding they were permitted to exploit it to recoup their costs over a specified period. Thereafter every ten days an auction was held at which contractors submitted bids {nyusatsu) of how much levy (unjo) they would pay and mining rights were awarded to the highest bidder.^ In 1603 Ieyasu appointed Okubo Nagayasu X^^^"^ (1545-1613), who was, like his predecessor, an experienced and able administrator trusted by the Shogun, to take charge of operations on Sado. Nagayasu was responsible for technological and administrative improvements which considerably increased the profitability of the mines for the government. He established his office (Jin^ya WM. or Bugyosho ^tfpff) at Aikawa and revised the way in which the mines were managed. Although the unjo nyusatsu procedure had succeeded in stimulating the opening of new mines and productivity in the short term, it did not provide resources for the necessary long-term investment and so Nagayasu initiated the nmake ^^ (division of profits) system. The precise details of this altered over the years but its basic principle was that in jikiyama Ef-U (literally 'directly controlled mountains' (i.e. mines)) the Bugyosho provided funds for major works such as tunnelling, improving drainage and introducing new technology. It also paid for materials and equipment. The ore extracted from the mine was divided with the mine operator (j/amanushi i^i). The rate varied over the course of the Edo Period (1600-1868) starting at around forty per cent for the Bugyosho but as the Shogunate took on more and more aspects of running the mines this increased to more than sixty per cent. Other privately run mines {jibunyama U ^>UJ ), although under the ultimate control of the Bugyosho, were left to the day-to-day control of entrepreneurs who did not receive subsidies and who paid a straight levy on the ore they extracted. The Archimedes screw pump (suijorin i^^M) was first introduced in 1618, in an effort to tackle the perennial problem of flooding, and this heralded Sado's most productive period. During the 1620s it is estimated that an average of six to nine of gold and silver were extracted annually. But the boom was short-lived and by the middle of the seventeenth century the industry was in decline for a variety of reasons. Following the closing of the country to all but a handful of foreigners in 1639 overseas trade was drastically reduced and the value of gold and silver declined. The search for new ore deposits led to the digging of ever deeper tunnels. The expense of constructing them and keeping them from flooding caused costs to climb steeply and the Shogunate's annual revenue decreased, dropping from over 6,000 kan (22,500 kilogrammes)** of gold and silver in the 1620s to less than 1,000 kan (3,750 kilogrammes) by the end of the 1660s, In 1651 the government ordered the closure of the less profitable mines. When a new Bugyo Hagiwara Shigehide ^i^iiriM^ was appointed in 1690 he made a concerted effort to revive operations and invested heavily, including the excavation of a new thousand-yard- long drainage tunnel. As a result of his efforts revenue rose to 2,000 kan (7,500 kilogrammes) per annum in the first decade of the eighteenth century but thereafter the decline continued with the Tokugawa Bakufu reluctant to risk further spending, and by the 1730S its income from the mines had fallen to 300 kan (1,125 kilogrammes). It was not until the beginning of the Period that matters improved. Following the end of its seclusion policy the Shogunate dispatched an Enghsh mining expert, Erasmus Gower (1830-1903), to Sado to make a thorough survey of the mines. He arrived in January 1868, just as the Tokugawa regime was collapsing, and had to return to Tokyo almost immediately. In 1869 the new government took over the mines on Sado and sent Gower back to the island to resume his work. He remained until 1872 and with other foreign experts arranged for the provision of more modern equipment in the hopes of resuscitating the mining industry. In 1886 the Meiji government decided to introduce non-convertible paper currency. The need to produce silver coinage to exchange for this brought about a revival in Sado's fortunes and in that year its mines produced 36 kan of gold and 938 kan of silver. Ten years later the island's mining operations, together with those at Ikuno in Hyogo Prefecture and the Osaka Refinery were sold to the Mitsubishi Company for 1,730,000 yen. Mining continued on a small scale until the last mine finally closed in July 1985. It has since been reopened as a museum.

SADO MINING SCROLLS IN GENERAL The first of the mining scrolls is believed to be Shokasegikata emakf il^:^^#, commissioned by the Sado bugyo during the Kyoho Period (1716-1736). It underwent major revisions in 1759 and again in 1819 to refiect various changes in the technology employed or in the administration of the mines. Kinzan emaki were produced in some numbers during the Edo Period and depict many different mine sites around Japan.^ A survey of scrolls representing the Sado mines carried out by the TEM Research Institute as part of their preparations for constructing a scale model of the Aikawa mine^ found seventeen scrolls on Sado itself and twenty- three elsewhere in Japan. Ford and Brown^*' mention three scrolls in the United Kingdom, although, interestingly, not the British Library's, and examples are known in other European collections.^^ Although scrolls depicting other sites are found, the majority represent mines on Sado. The Sado scrolls were painted by artists, known as ezushi ^[Uffi, who were employed by the Shogunate office in charge of the island's mining activities, the Bugyosho. The first of these was Yamao Kakken \hMM%i , who was appointed by Hagiwara Yoshiharu V^ Sado bugyo from 1732 to 1736. Their primary purpose was to provide an

132 accurate representation of the complicated workings of the mines, both below and above ground for the benefit of the hugyo and other officials appointed by Edo every few years. Since their function was to explain the current state of operations it was necessary to have a new scroll prepared when changes in working methods or adminstrative procedures occurred. Scrolls would also have been sent back to the government in Edo to keep it informed of developments. The bugyo and other officials may have received copies of the scrolls as keepsakes when their term of appointment came to an end. What better memento of one's tour of duty as an official on Sado than to have a copy of one of the scrolls made.' It is interesting that the number of scrolls seems to have proliferated in inverse proportion to the productivity of the mines. One explanation is that their role came to be less didactic and more artistic. Quite apart from differences in content there is an enormous variety in the artistic merit of the extant scrolls. Some are simple pen and ink drawings on paper, others, like the British Library's, are skilfully and elaborately painted on silk. Clearly the latter must be seen as art objects rather than simple text books. Following the opening of the country to foreigners in the 1850s the authorities realized that a copy of one of these illustrated manuscripts would make an excellent gift for a visiting expert or notable, which is how many have passed into Western collections. Unfortunately there is no record of Siebold's ever having visited Sado so the likelihood is that he acquired his from a bookdealer on the mainland.

THE BRITISH LIBRARY S SCROLLS In a letter dated 20 June 1868 to Thomas Watts of the British Museum which describes the contents of his father's collection, Alexander von Siebold mentions a number of works in a category entitled 'Gold, silver and copper coins' which included the three Sado mining scrolls. These were added to the British Museum oriental manuscript collection and assigned the numbers Or. 918, Or. 919 and Or. 920. In 1973 the Museum's Department of Oriental Printed Books and Manuscripts was transferred to the newly established British Library and the scrolls now form part of the Japanese collection of the Library's Oriental and India Office Collections. The three scrolls, in fact, constitute a single manuscript which shows in minute detail the processes involved in the extraction and processing of the ore as well as the administrative and commercial activities associated with the mines. The British Library manuscript does not have one overall title; instead each of the three scrolls has a gold paper slip glued to its outer cover bearing an individual title which is taken from the caption accompanying the first scene on each scroll. Thus Or. 918 is Kobandokoro ntte sujigane tamabuki no zu 'b'¥\rfi—"r^^'^\^X~^^ (Illustrations showing the of gold balls in the mint). Or. 919 is Dotokoya^''^ no zu ^0f.M^^ (Illustra- tions of the copper smeltery) and Or. 920 is Kinginzan shikinaioka kasegikata no zu :^0Uj|feF^[^^:^5:iE](Illustrations of working methods inside the gold and silver mines). Comparison with other known scrolls of this type reveals that Or. 920 is actually the first 133 part of the manuscript. Or. 919 the second and Or. 918 the third. Clearly the member ot the British Museum staff who assigned the manuscript numbers well over a century ago was not aware of this nor, it would seem, was von Siebold since two of the scrolls still bear labels with the numbers they had in his collection. Or. 919 is 398 and Or. 920 is 397. Or. 918 has lost its label at some point so there is no way of knowing whether it was 396 or 399. Given that the sequence of the scrolls as indicated by their shelfmarks is incorrect, it seems reasonable to apply the title Kinginzan shikinaioka kasegikata no zu, the title of Or. 920, to the whole manuscript. Each scroll is made up of a number of silk panels, backed with Japanese mica- embossed paper and glued together. A white silk edging has been added to the top and bottom of each panel. A brown brocade cover with gold floral designs is attached at the beginning of the scroll while the end of each is fixed to a wooden roller. The width of the scrolls is consistently 23.5 centimetres for the painted portion with a further i centimetre edge top and bottom making a total width of 25.5 centimetres. However, it is clear that the scrolls have been trimmed since in some places the captions to the illustrations are cut off at the edge. The length of the individual panels varies considerably, averaging around 120 centimetres. Or. 920 has seven panels while Or. 919 and Or. 918 each have six. The overall lengths are 822 centimetres for Or. 920, 714 centimetres for Or. 919 and 752 centimetres for Or. 918, giving a grand total of 2,288 centimetres. The precise dating of Kinzan emaki is extremely difficult. Only a very few examples bear any date or indication of the artist. Attempts at identification have to be made on the basis of the internal evidence provided by the scenes depicted, taking into account developments in mining technology, buildings, changes in the management and administration of the mines and even the costumes and hairstyles of the people shown. For example, the Yoseseriba ^)^^ , where the grading of the ore took place, was constructed in 1759. Since this building is depicted in the second of the British Library's scrolls, Dotokoya no zu, the manuscript must have been drawn after this date. This is supported by the type of lighting shown in the mines: the distinctive iron oil lamps, know'n as tsuritomoshi (^It), replaced the earlier wooden torches during the 1750s. Another useful aid to dating is the type of bellows being used to heat the furnaces in the smithies and smelting workshops. Originally each furnace had its own set of bellows ifuigo W ) alongside it operated by a man called a fuigosashi |i^ . This arrangement was called kataftiigo W^% ('one-side-bellows'). From the Kansei Era (1789-1801) the decline in productivity and a resultant need to reduce the work force led to the introduction of a system called ryobuchifuigo \^WM ('both-sides-bellows') where one set of bellows was placed between two furnaces, thus halving the number of operatives required. The illustrations in the Library's scrolls show only katafuigo and it is therefore reasonable to assume that the scroll depicts activities before 1789. It can thus be concluded with reasonable confidence that the manuscript represents the appearance of the mines between 1759 and 1789. It does not, however, necessarily follow that that is when the scrolls were painted. Although the main variants of the genre can be roughly dated from

134 their content as noted above, copies of each of them went on being produced for considerable periods. One of the few examples o^ Sado kinzan emaki for which an artist can be identified is in the possession of Aikawa Local Museum (Aikawa Kyodo Hakubutsukan ^lJf on Sado. Entitled Kinginzan emaki it is known to be the work of Ishii Natsumi (1782-1848). Its overall length is almost exactly the same as the British Library manuscript although it is divided into two rather than three scrolls. Its content is almost identical and it is very similar in artistic style. Although no external corroborating proof exists that our manuscript is also by Ishii Natsumi, it is certainly a tantalizing possibility. Even if one rejects this hypothesis it is still highly likely that the two manuscripts were painted at roughly the same period: around 1810 to 1820.

CONTENT OF THE SCROLLS The first portion of the manuscript, scroll Or. 920 Kinginzan shiktnaioka kasegikata no zu, shows the activities underground in the mine itself, from the entrance, known as the kama no kuchi ^p or J mouth of the cauldron', over which hangs an invocatory plaque to the mountain deity Oyamatsumi;^Uj^lt, on one side of the mountain to the exit on the other. The first figures encountered are mine workers called niage horiko ^^^T- ,1^ whose job it was to bring the mined ore up to the surface. Clad in cotton jackets and leggings (kyahan mm), with straw sandals (maraji ^^ ) and a cloth turban, they can be seen ascending and descending by means of ladders made from logs into which steps have been cut (marukibashigo ^^^^) or, where the gradient was too steep for this, by means of short wooden steps inserted into the rock on either side of the passage (uchikae t^m ). They carry small lamps (tsuri ^ or tsuritomoshi mm) made of an iron dish to hold the oil attached to a long iron handle. The lamps could also be wedged into crevices in the rock to provide a fixed light for the miners. On their backs the mage horiko carry straw sacks (kamasu "A) which, when full of ore, would weigh 20 kilogrammes and were checked by an inspector who sat near the exit to keep a tally of production. Further along the miners themselves, the kanahori daiku ^^^x , are to be seen, seated cross-legged and bare-chested on small straw mats. They use long-handled tongs called uedabashi ±03"^ to hold the iron wedges or 'gads' (tagane m ) which were hammered into the rock to split it open. The miners worked in pairs in eight-hour shifts and while one laboured at the rock face his partner would take his meal and rest breaks. The conditions must have been grim indeed. The average dimensions of the tunnels were 2.4 metres in height and 1.8 metres in width and it has been calculated that a miner would be able to advance between 9 and 19 centimetres per day depending on the hardness of the rock. The number of miners varied considerably depending on the productivity of the mines: during the boom of the early 1600s there were up to 450 but this had dropped to ninety by the middle of the next century. At the same time, elsewhere in the mine, the search for new seams is going on. A group of officials including xheyamanushi, the mine operator, the surveyor called xhtfuriganeshi ll^teSifi (literally the 'master of wielding the ruler') and the mining inspectors are discussing the best location for a new tunnel while the contractor, the kengiri ukenushi fHi^J.=fli , who will actually do the work sits and waits. In a small concession to health and safety at work they wear protective pads (lehen JM^Jl), made from twisted paper string, tied over their cloth turbans. Flooding posed a perpetual threat and control of water in the mines was a prime concern. Various methods of drainage were devised over the years. The most basic was removing the water by a hand-operated system of buckets and pulleys, a process called teguri mizukae ^W^^ , which transferred the water into the long drainage channels through which it flowed out of the mine. The Archimedes screw was also widely used on Sado from the mid seventeenth century but does not appear in use inside the mine on the British Library scroll. Another problem was ventilation. In an attempt to remove the foul air, filled with smoke from the lamps and fires, dust, and the smell of human waste, mechanical fans of a type known as a 'blow George' were introduced to the mines. Called kazamawashi toumi IEl[Hl/tK ('wind-turning Chinese fans') they were based on winnowing fans introduced from China at the end of the seventeenth century. They consisted of wooden drums with a handle which, when turned, moved blades inside the drum to draw air through a series of wooden ducts from the outside into the depths of the mine (Plate Va). Another important category of workers shown on the scroll are the carpenters or yamadome daiku LJJ^^X responsible for constructing the wooden supports (yamadome), platforms, steps and ladders used in the mines. They were assisted by the tetsudai horiko ^•fE#^ who carried the pit props down to the tunnels where they were required and assembled them, and the choba horiko T^^^ who removed the waste rubble. Approximately half the first scroll is taken up with the activities underground but then the focus shifts to surface operations. Outside the mine the first building encountered is a small hut called the Yokohikiba 1^^1*i where the bags of ore (kusari m )^^ brought up by the mage horiko were weighed while nearby the carpenters are busy fashioning pit props. A little further on is one of the kajigoya m'i^'hM. , the workshops for the blacksmiths (kajtgata m^:^) who made the tools, wedges, nails and so on used in the mines. Making the gads or tagane was a major part of their work since each one would last only two or three shifts before it broke or became too blunt. Teams of smiths worked day and night to keep up with demand. Mention has already been made of the bellows used to keep the furnaces in the smithy at the right temperature. Although the British Library scroll shows them simply as square shapes behind each furnace, other sources show that they were rectangular airtight boxes made of thin wooden boards. Inside was a piston with a long handle and pushing this backwards and forwards created the draught. The pistons were covered in badger skin and the demand was such that badgers were driven to extinction on Sado. Depending on the size of the bellows they required one or two operators (fuigosashi or fuigosasht horiko) and it is interesting to note that the scroll shows them being worked by young boys. This was just one of the jobs that 136 children performed around the mines until from the age of fourteen or fifteen the boys were allowed to work inside the mine itself. Women also figure frequently in the scenes. One of their main tasks was to remove the waste material from the ore to leave behind the portion containing the gold and silver, a process known as ishieri ;>. This was done in a building called the tateba mm which contained a large water-filled tank. The women, known as ishierime ^^:^, placed the ore in sieves and washed it in the water, then chipped away the unwanted lode with hammers under the watchful eye of supervisors (Plate Vb). The next section of the scroll shows the bustling interior of the Yottsudome bansho El'y§#^, the administrative centre for the mine. In one part of the building are storerooms: the sumi nay a ^^g where supplies of charcoal for the blacksmiths' furnaces were kept and the tomegi naya ^i^'^m. for the pit props and other wood used in the mines. In another is the abura bansho^^W\ or 'oil office' responsible for providing the oil for the miners' lamps and other items such as the metal gads. There is the accounts office where the book-keeper is shown with his abacus, carefully recording all the transactions, hours worked, wages to be paid and so forth, while in the next room there is a meeting of the managers and senior administrators of the mine. The Yottsudome bansho was also the place where the ore was graded for sale to the smelters. At one end of the building one of the ishierime is delivering a consignment of ore which is examined by officials, placed into sacks and labelled according to its quality. In an annexe at the other end of the building the smelting contractors or kaishi «;& are examining samples of the ore which they will then bid for at auction. This process was known as niuri ^^ and took place every ten days. To ensure that no potentially valuable were overlooked, the material discarded by the ishierime was checked over in the dobegoya,^^ small sheds consisting of a pit dug into the ground with a sloping thatched roof constructed over them. Here men called dobehon sieved the waste ore they had collected and sold any small quantities of gold or silver that they found to the smelters or direct to the Bugyosho. The final stage in the administrative process shown in the scroll is the division of the proceeds between the mine operator, the contractors and the Bugyosho, the mwake referred to above. Once this was completed the ore was sewn into sacks and carried out through the mine gate called the Ai noyamaguchi TPILUP where it was loaded onto oxen to be carried off to the smelting works. The first scroll ends with a scene of life in the nearby town of Aikawa where the mine workers lived. There are off-duty miners drinking sake in the wineshop, shops selling torches, baskets and other equipment, a fish- seller and children playing in the street - perhaps it was their day off or perhaps they were still too young to be employed at the mine. From its origins as a small fishing village of around fifty to sixty houses with a population of two to three hundred at the beginning of the seventeenth century Aikawa grew rapidly as the mines developed. Estimates of the total population vary widely but it is believed that by 1629 when the boom was at its height there were 3,790 houses with a total of more than 200,000 inhabitants.^'^ The second scroll, Or. 919 Dotokoya no zu, opens with the processing of copper in the 137 copper smeltery or Dotokoya m^Mm . The order of the elements in the extant scrolls varies and in many the Dotokoya comes at the end of the manuscript after the scenes of gold and silver smelting with which it shared many common features. A description of these processes is given below in the section dealing with the Yoseseriba}'' From the copper works we move to the coast of Sado where we see the extraction of alluvial gold and silver from the on the beaches (hamanagashi Wik ) by means of a technique called ' sluicing' or nekonagashi t^ ^ ij^ L. Sand was collected, placed in sieves or on w oven mats and sluiced in a stream or water-channel to wash away the lighter sand, leaving behind any traces of gold or silver. Where no running water was available a pit was dug on the beach and filled with water. An Archimedes screw is shown drawing the water up out of the pit to form a flow that could be used for sluicing. The material was further processed in a building called the Hamanagashi no seriba using technology discussed below. After this brief excursion outdoors the action moves back to the buildings near the mine and into the Yoseseriba ^^^. This was an extremely busy place echoing with the sounds of hammers and grindstones and forming the centre for the concentration of the ore won from the mines. As was noted above there were other seriba where other material was dealt with - for example, the copper and the alluvial gold and silver - but the complex series of processes was fundamentally the same in each. At the entrance and exit of the Yoseseriba was a checkpoint where the passes of workers and officials were inspected. On the way out people were subjected to extremely thorough body searches to ensure that they had no scraps of ore or refined gold or silver concealed in their clothing or about their persons! The various stages by which the precious metals were won from the ore were known collectively as konashi fuf& and the first of these, pulverization or ishihataki ^fD , took place in an area called the Kanaba ^^ . On the British Library scroll this contains four kanabaishi ^ii-^ - large blocks of black stone, either andesite or basalt, each with a ^ hollow dug into the top of it. Into this ore was placed and then pounded with an iron hammer weighing four to five kilogrammes. On average a worker would pulverize thirty to thirty-five kilogrammes of ore per day. The pulverized ore was then sieved using an uma no ofurui M(OW^U (literally 'horsetail sieve'); by means of a series of sieves of different fineness it was possible to separate the ore into various constituents: ore containing principally gold (mizusuji TKISE ) and that containing principally silver (ymrunono ik^ ). The rest of the material which could not be separated easily in this way was sent for further processing. This was achieved by grinding or usuhiki M^. The grindstones (ishiusu ^^) consisted of two parts: a lower stone which stood in a shallow trough of water and an upper stone to which a long wooden handle was attached. The ore was placed into a hole in the top of the upper stone, water was added and the stone was rotated by two or three workers pushing and pulling on the handle. This process reduced the ore to the consistency of mud which was then scooped out of the trough and placed into water- filled tubs called lateoke

(a) View inside the mine showing in the centre the *blow George' fan. At the bottom left workers armed with buckets are busy removing water. Or. 920 (detail)

g^^

(b) Above: the blacksmiths' workshop; below: the ishierime washing the ore and removing the unwanted lode. Or. 920 (detail) PLATE VI

(a) Grinding and sieving the ore. To the left can be seen the wooden troughs for nekonagashi with the nekobaba's empty platform. Or. 919 (detail)

(b) The cementation process, showing (bottom) the large wooden shioawasebune where powdered gold was mixed with salt and (top) the nagakamado where cones of the resulting mixture were baked. Or. 918 (detail) By stirring the sludge the gravimetric principle was used to separate the different minerals, the heavier ones sinking to the bottom, the lighter ones remaining suspended in the water. Next a slightly concave wooden jig (yuri-ita ^tS) was used to agitate the precipitate in a vat of water so that the heavier particles, notably the gold and silver, stayed on the edge of the pan while the lighter ones were washed away. The water containing the finer particles was put into a large wooden tank called an uwabune .t^. To extract the very small particles this water was subjected to the process oi' nekonagashi already referred to. This involved a series of wooden troughs 3.5 metres long and 24 centimetres wide which were constructed on a gradient of i in 12 and into each of which a length of cotton cloth was laid. Water from the uwabune flowed over them and the gold and silver were precipitated onto the cloth which was then rinsed in a tub and sieved once more. Much of this work was done by women and many scrolls, although sadly not the British Library's, show a female supervisor sitting on a platform over the headtank. This was the nekobaba and she is thought to be the origin of the Japanese expression nekobaba 0 kimeru which means to 'embezzle', which does not say much for her reputation for honesty! ^^ Even the richest ore contained only small amounts of silver and gold, along with a large proportion of copper, and so it was necessary to concentrate the metal. The various steps in this process were carried out in a building called the Yosedokoya ^^M whose most striking feature was the tall towers of the blast-furnaces or otoko ±1^ . These were wooden constructions with a hearth (ro '^) and set of box-bellows at the bottom and a chimney (kemuridashiMl^ L ) at the top. Each furnace was worked by a smelter called ^fukidaiku P^:;'VXassisted by a man or boy to operate the bellows. Firstly, the auriferous copper was smelted with which attracted the silver to itself. Next, the gold/silver/lead alloy was taken to an area called the Haifukidoko mX^^ where it was subjected to or 'air scorification' in an ash-lined iron dish or 'cupel' so that the lead oxidized, leaving behind the molten silver and any gold it might contain. This principle of or haifuktho mXii: was introduced from Korea during the early part of the sixteenth century and is first recorded in use in Japan at mines in Iwami in 1533. The gold and silver could then be separated by melting them repeatedly with sulphur which combined with the silver to form silver sulphide and then removing the latter to leave behind the gold, termed sujigane %^}^ The second of the British Library scrolls ends at this point and the third. Or. 918 Kobandokoro nite sujigane tamabuki no zu, is concerned with the production of coinage using the obtained by the operations shown in the first two. The Kobandokoro^^ was the mint in which small coins called koban /J^^ij ^^ were produced. The koban, which had a face value of one ryo m , was introduced by Tokugawa Ieyasu at the beginning of the seventeenth century and remained in general circulation throughout the Edo Period. It consisted of a thin piece of gold cut in an oval measuring 3.5 X 7.5 centimetres and weighing 4.76 momme (17.85 grammes).^^ Initially a koban was 84.29 per cent gold so that a ryo was equal to 15 grammes of gold. However, over the next two and a half centuries the koban were recalled and reissued nine times with their 139 gold content being progressively reduced until by the i86os they weighed just 3.3 grammes and were only 56 per cent gold. Before 1622 gold from Sado's mines was taken to Edo for minting but in that year Goto Shozaburo Mitsutsugu 'ikMf^^^Mit'lk (1571-1625) was ordered to set up a mint on the island. The Goto family had long enjoyed a reputation as master metalworkers and in 1594 Tokugawa Ieyasu had asked Mitsutsugu to begin minting coins. Once Ieyasu became Shogun he established a number of gold and silver mints, called kinza ^J^ and ginza ^^ respectively^^ as part of his strategy to control the nation's economic and commercial life. Successive generations of the Goto family held the office of gokingin aratameyaku M<^M'^'^ (Inspector of Gold and Silver Mints) or X^ltr gokin aratameyaku i^^efe^K (Inspector of Gold Mints) until 1869 when the office was abolished. At the beginning of the third scroll the sujigane is being cut up, washed, melted over a charcoal fire and dried. It is then weighed and wrapped in batches of 500 momme (1,875 grammes) by Kobandokoro officials under the scrutiny of Shogunate inspectors {metsuke g-fTJ-). The purity of the gold is then assessed and recorded. The gold obtained by the concentration described above was generally only 60-70 per cent pure at this stage and further refining was needed to render it suitable for the manufacture of coins. This was achieved using a cementation process called shioyaki S'^3^ (lit. 'roasting with salt') which involved the chloridization of the silver still contained in the sujtgane. The sujtgane was melted and then scooped onto an iron plate, allowed to cool and ground to a powder. The scroll shows suspended above the SaikindokoW^^ , as the area in which this took place was called, a criss-cross construction of wooden planks called a senryodana ^jSjffl or 'thousand ryo lattice'. This was designed to trap any gold dust mixed with the smoke from the smelting so that it could be recycled. The powdered gold was mixed with salt in a large wooden box called a shtoawasebune i^^^^ SLiid the mixture of salt and gold powder was shaped into cones using a tubular wooden mould. Each cone was placed on an earthenware dish and a small earthenware lid was laid on top before they were put in rows in a long shallow pit called a nagakamado W^ or 'long oven' (Plate Vlb).^^ The pit was filled with wood and when this was burnt slowly over a period of seven or eight hours, the heat caused the silver particles mixed with the gold to react with the salt to form silver chloride. The cones were carefully removed from the pit with long-handled tongs by workers wearing tall wooden pattens (takageta i^T-li), since the earth was still very hot, and tipped into wooden tubs. Next the tubs were stood in a water-filled trough and workers trod the mixture until the silver chloride dispersed. The gold retrieved in this manner was now called yakigane or yakikin *^^ (literally 'roast gold'). It was taken to an area called the MomiganedokoM^^^ or 'Gold-rubbing workshop' which was also equipped with large water tanks. Into the tanks were placed small buckets and on top of them baskets containing the yakigane. By rubbing the yakigane under warm water any remaining salt or waste material was removed. The water from the tanks was sieved to ensure that any small particles of useful minerals were recovered for recycling. In the final stage of the refining process the yakigane, now termed momigane ^^ or 140 ^. /. Scenes in the Kobandokoro where the gold is hammered into strips and polished with salt under the watchful eye of an inspector seated on the platform in the centre. Or. 918 (detail)

'rubbed gold', was melted once more to form balls of gold called yosegane ^^ or 'gathered gold'. This was the basic material needed for minting the koban (fig. i). Since pure gold was too soft for the manufacture of coinage a proportion of silver was added to produce an alloy. As was explained above, the exact ratio varied over the course of the Edo Period but initially 0.75 momme (2.8 grammes) of silver was added to 4 momme (15 grammes) of gold. The alloy was melted and poured into shallow, oblong bamboo moulds called takenagashidai ^'(kl--^. Once the metal had cooled it was removed from the mould as a metal strip called a nobegane or nobekin 5£^ ('elongated gold') 90 centimetres in length and 6 centimetres in width. This was hammered to the required thickness of 1.6 millimetres, polished with salt and given a final wash before being sent to the office run by the Goto family, the Goto Yakusho '^B'<^f^ or Gotoza 'ikMf^..Her e each strip was stamped with a mark called a gokuin ^^\\ and cut into two. One half was assessed for quality while the other was tested to see whether it contained copper. This was done by making two small indentations in the middle of the half-strip and touching these with a stick of burning hinoki (Japanese cypress) wood, the colour of the resulting smoke indicating the presence or absence of copper. If it met the prescribed standard it was certified with another stamp. Next the strips were sent back to the Kobandokoro where they were each cut into 120 small sections measuring 3 x 1.5 centimetres. Each of these was then worked into the distinctive elliptical shape of the koban by repeated heating and hammering. The necessary stamps were applied by the Gotoza inspectors and finally the coins were dipped in a gold-coloured dye to disguise the whiteish shade caused by the silver in the alloy. Once polished they were sent to the Bugyosho to be stored in the strongroom or Gokinzo 'i^-^M ready for shipment to Edo. At the height of 141 Sado's productivity two or three consignments of coins and refined gold, silver and copper were dispatched to Edo each year but from 1754 this was reduced to one. The cargo was taken by ship from Aikawa to Izumosaki where it was loaded onto horses for the overland journey to the Shogunate accounting office, the Kanjo Bugyosho iti^^tTfiff. As has been discussed elsewhere in this article the quantity of precious metals produced by the mines of Sado fluctuated considerably. The story was fundamentally one of a decline from the boom period of the early decades of the seventeenth century, interspersed with episodes of increased production following the introduction of new technology or the discovery of new seams. The total value of gold and silver mined was as high as 152,629 ryo in 1623 but had fallen to just 9,084 ryo by 1866.^^ The final scene of this third scroll is a rather naively painted panoramic view of the gold mines and sites for gold-panning to be found in the mountains of Nishimikawa in the south of the island. The British Library's three Sado kinzan emaki give a detailed insight into the highly complex world of mining in pre-modern Japan. They show in an attractive and lively manner what went on both below and above ground and provide a wealth of information for metallurgists and scientists interested in the techniques, processes and accoutrements of traditional mining, whilst at the same time giving the more general student fascinating images of the daily life and activities of the people of Sado in their attempts to exploit their island's riches.

The author gratefully acknowledges the assistance of This system is similar to that wideiy used in Dr Paul Craddock, Head of the Metals Section, Europe at the time and may have been influenced Department of Scientific Research, British Museum, by contacts with the Portuguese and Spanish. in clarifying some of the more technical aspects of One kan = 3.75 kg. traditional metallurgical and mining techniques. Nihon Gakushiin g^^±P^ , Meiji~zen Nihon 1 Von Siebold (1796-1866) had spent the years kog}'d gijutsu hattatsushi a^^pfii P^SE^^ii^ 1823-9 as doctor for the Dutch East India (Tokyo, 1958), p. 212. Company at Deshima. However, almost all the Nihon no kozan bunka: ezu ga kataru kurashi to Siebold items now in the British Library were gijutsu \ii^(Dm.\h%{t:^mtm^M^h\. t&M , collected during his second sojourn in Japan a catalogue of an exhibition held at the National between 1859 and 1862. For an overview of the Science Museum in Tokyo in 1996, provides Siebold collection, see Yu-Ying Brown, 'The illustrations of scrolls depicting more than Von Siebold collection in The British Library', twenty different mines. British Library Journal, i (1975)^ PP- 163-70, and TEM Research Institute (ed.), Zusetsu Sado ii (1976), pp. 38-55. A revised version in kinzan [^Ifeteffi^UJ (Tokyo, 1985). Japanese appeared in Shiiboruto no mita Nikon 10 Trevor D. Ford and Ivor J. Brown, 'Early Gold v—^Vu h(/:)EfcP^ (Tokyo: NHK Books, Mining in Japan: the Sado Scrolls', Bulletin of 1998). the Peak District Mines Historical Society, xii, no. 2 The word kinzan (also sometimes read as 6 (Winter 1995). kanayama) literally means 'gold mountain(s)'. II For a description of Sado scrolls in German 3 Among those who were exiled to the island were collections see H. Winkelmann, 'Das Sado- the Emperor Juntoku (i 197-1242), the Buddhist Goldbergwerk auf Japanischen Rollbildern', Der priest Nichiren (1222-82) and the dramatist Anschnitt, ix, no. 4 (i957)- Zeami (1363-1443). 12 These characters may also be read as Doshoya. 4 The story is to be found in vol. 26, no. 15 of 13 The word horiko, derived from the verb horn Konjaku monogatari. meaning 'to dig', was a generic term for all the 142 workers involved in the mining process, not on the title slip of the British Library's scroll simply those who mined the ore. They were clearly shows that it was intended to be divided into various categories which will be pronounced as Kobandokoro. introduced below. 21 The other gold coin widely used during the Edo 14 In the scrolls this word is employed rather than Period was the oban ::*cf-iJ which was similar in the more usual koseki ^X.^ . shape to the koban but much larger, measuring 15 Dobe is a dialect word for 'mud'. up to 10 X 15 cm. and weighing 165 g., with a face 16 Hasegawa Riheiji :l©jll*ij^i^ , Sado kingin- value of 10 ryo. zanshi no kenkyu fe^^lSlil^cofiJf^ (Tokyo, 22 One momme — 3.75 g. 1991), pp. 123-30. 23 The fashionable Tokyo district of Ginza owes its 17 For a very detailed explanation of the traditional name to the silver mint that existed in the area Japanese techniques for smelting copper, see from 1612 to 1800. Cyril Stanley Smith (ed.), Kodo zuroku: illus- 24 The evidence of the scrolls supports William trated book on copper smelting (Norwalk, 1983). Gowland's widely accepted account of the 18 A less attractive etymology for the expression cementation process traditionally used in Japan refers to a cat's habit of burying its faeces. in which he states that the salt and gold were 19 In modern Japanese this expression, which may mixed together and simply placed on the furnace be literally translated 'rod-shaped gold', is used floor for cementation. This is significant since in to mean 'metal reinforcement' or to imply that a other parts of the world traditional methods of person has ' backbone'. In Japanese works cementation involved the use of closed vessels. dealing with traditional metallurgical techniques, See W. Gowland, ' Silver in Roman and Earlier however, it is normally equated with shizenkin or Times', Archaeologia, lxix (1917/18), pp. 121— 'natural gold'. 60. 20 This word may also read be as Kobansho or 25 For detailed figures of gold and silver production Kobanjo but a phonetic transcription {furigana) on Sado, see Hasegawa, op. cit. in n. 16.