On one page of Charles Lang Freer's papers a pair of Japanese folding screens has three prices. There is a figure of eight thousand yen, a second figure of eight thousand eight hundred yen and a third figure of four thousand four hundred and seventy dollars. They occupy the same entry. On the very next lines three pottery vessels occupy space in the collection account without a purchase price at all. The handwriting and typewriting of a busy office have left a more discriminating account of how art moved than the single column headed “price” in many modern research notes can accommodate.
The museum visitor asking what a museum paid has a perfectly reasonable question. The first surprise in this instance is the identity of the buyer. It was Freer who bought the pair in 1909. The voucher belongs to the history of a private collection that became part of a national museum. A London example, by contrast, records a museum buying an already assembled collection from a private individual. The two events look much the same from the gallery floor. Each places Japanese art in a public collection. Follow the money and they belong to quite different histories of power, scholarship and choice.
Two screens and one line on the account

The Freer documents are preserved at the Smithsonian. In its transcription project for the 1909 art vouchers, page 748 provides a summary and page 797 a descriptive entry. The numbering is the digitization project's pagination within a much larger series, so each page also has its own archival identifier. The description gives the reader two sixfold screens, floral imagery and a gold ground. One screen carried hydrangeas and lilies, among other plants, while white peonies appeared on its companion. The account treats the pair as one acquisition. It would take two photographs to illustrate the entire pair, but only one row to record the negotiated price.
The pair illustrated in this article, now in Cleveland, makes that distinction visible particularly well. It is a separate work by Watanabe Shikō. Its photographs show how a viewer can look at one stretch of vegetation and gold at a time while knowing that the larger composition has a second half. Cleveland assigns each screen its own identifying suffix. The photographs supply a visual example of a type of possession that may travel in two physical parts and nevertheless be bought as one. They do not depict the pair acquired by Freer. Their collection numbers and the names in their credit lines prevent the two cases from drifting together.
The summary for Freer's pair records eight thousand yen and an additional ten percent commission. The arithmetic produces the second figure, eight thousand eight hundred yen. A dollar entry of 4,470 follows. The descriptive sheet also mentions freight and other expenses in abbreviated form. Neither page, considered on its own, provides an itemized freight bill for just these screens. A modern reader can verify the addition of the commission with confidence and still have a sensible question about the other charges. The surviving figures permit an answer more informative than either a single unexplained price or a claim that the price is unknowable.
The paper also preserves two moments in the life of the acquisition. The object description dates the encounter to a summer journey, while the voucher belongs to December. A catalogue giving the month of settlement tells the truth about an account. A travel chronology giving the season tells the truth about the collector's movement. Putting the two together discloses an interval during which belongings, bills and instructions might move by different routes. It does not establish the exact day that the screens were packed, left Japan, arrived in America or first stood open in a gallery. Those are all legitimate questions for other documents.
A bank transfer across the Pacific
The Peoples State Bank in Detroit enters this story on another sheet. The transcription of page 788 records a cable to Samurai Shokai in Yokohama on December 21, 1909. The amount transmitted was 48,040.50 yen. The document, dated December 24, includes the cost of transmission and acknowledges 24,030.89 dollars in settlement. The signatures and marks associated with checking and payment give the document a practical function. It helped close an account. Reading that number into the screen's object record as its purchase price would be a remarkable mistake, even though the banker and the object sheet both name the Yokohama firm and belong to the same month.
The transfer is much larger than the sum entered against the screens. That difference is of positive historical interest. The bank and the collector needed a means to settle an aggregate obligation; an object historian wants to recover a particular purchase. A reliable common reference joins the papers without making the transactions described on them identical. The document naming a recipient of money can confirm an intermediary's role. It cannot, by virtue of naming that intermediary, establish that every object in the account had previously belonged to that firm. In the screens entry the named seller and the business through which the transaction was conducted have different roles.
There is a tempting calculation here. Dividing amounts in yen by entries in dollars might appear to yield a historical exchange rate with considerable precision. Before crediting such a figure to a particular day, one would need to know what was included on each side. The bank sheet includes transmission costs. The summary and the item sheet use their available space differently; one explicitly adds commission and the other also mentions freight. Rounding to whole dollars on an object sheet can change an apparent rate again. The ratios can be calculated. A general statement about the market for foreign exchange in December would need more evidence than the resulting decimal places provide.
The name of the seller offers a quieter example of the same discipline. One public transcription spells it Tanaka and the descriptive page renders it Tanka. Both identify Samurai Shokai and the transaction for the floral pair. The Smithsonian's transcription instructions ask contributors to preserve spellings rather than silently modernizing them. For finding related papers it is worth searching both forms. For quoting a document, it is worth identifying which one supplies the spelling. A useful index can connect people across variants without erasing evidence of how their names reached the office that prepared the account.
What the zero would conceal

On page 748 the objects following the pair are three pear-shaped pottery water bottles associated with Owari. Each is described separately; all three are identified as gifts from Mr. T. Hara. The compiler distinguishes the shape of their necks and repeats the status of the transaction, sometimes by a ditto sign. A database that required a numerical amount might convert those marks to zero. The next analyst could then average them into a table of purchase prices for Japanese ceramics. The computation would be exact and the historical conclusion meaningless. The collector had received gifts. There was no bargain price of zero yen to compare with a seller's price for another vessel.
The water bottles also test the scope of a heading. A page continued under a heading for purchases in Japan can contain individual entries for gifts. Headings organize papers; individual lines can qualify them. Conversely, a scan that cuts off the heading may obscure the currency or the party through whom a group of purchases passed. Recovering a line in context requires looking both above and below it. The case of the three water bottles is especially instructive because the repeated term is meaningful information. An empty space, a ditto mark referring to a gift and a figure that cannot be read deserve three different treatments in a transcription.
Cleveland's tea bowl, accession 1983.158, provides a comparison closer to a label a visitor might actually encounter. Its credit names Iwao and Takako Setsu and honours Dr. and Mrs. Sherman E. Lee. Those are distinct relationships to the occasion of acquisition. The words do not reveal the price the donors may once have paid, nor do they purport to assess the bowl's monetary worth. The bowl is dated to about 1600. Centuries separate that approximate date of manufacture from the occasion recorded by the museum credit. Both belong to the bowl's biography, and neither supplies the missing terms of an earlier exchange.

The three Owari bottles and the Cleveland bowl thus illuminate the same archival problem from opposite ends. In one case the acquisition office records the nature of a transaction while describing several objects that arrived in the same period. In the other the museum credit publicly preserves the donor and the people being honoured. An economic history that omits gifts misses a means by which collections were formed. An economic history that assigns gifts a fictional price also misses what the records preserve. A reader can investigate the social meaning of an object changing hands even where the papers offer no numerical value at all.
The London collection bought for three thousand pounds
In Britain there is a different kind of answer to the question in this article's title. In 1881 the British Museum purchased much of the Japanese art collection assembled by William Anderson. Japan's National Diet Library gives the price as three thousand pounds. Michelle Huang's archival research at the University of St Andrews points to the museum's original papers, including the proposal and internal advocacy for the acquisition. Such evidence locates the decision within the institution. The buyer was a museum, the seller had assembled a collection, and the recorded amount concerns a body of works. These three qualifications are part of the answer.
There is no good reason to divide three thousand pounds by a count of works and present the quotient as the cost of one surviving drawing. Counting raises questions even before one reaches valuation. A bound album can be treated as a single accession or described leaf by leaf. Several studies might be attached to one sheet. A summary written for trustees might group material differently from a printed catalogue arranged for scholars. An average can be useful for a carefully defined research question, but a drawing did not acquire a separately negotiated historical price simply because modern arithmetic can give it a share of a group total.
The acquisition of illustrated books from Anderson supplies a particularly good control. Miya Itabashi's research distinguishes the 1881 purchase from a later transaction in 1882 for 311 illustrated books, at 360 pounds. If the question concerns the books, neither the date nor the amount for the earlier purchase answers it. The seller's name is the same, as is the collecting institution, and the cultural field is closely connected. Those similarities make it all the easier to conflate the transactions in a lecture caption. They also show why an accurate footnote should identify what was transferred and the occasion on which it changed hands.
Anderson was also a writer and cataloguer. The ability to describe a work affected how it could be seen in a public institution, but an acquired catalogue, a subsequent descriptive publication and the works described in it are separate entities. A historian following one Japanese drawing into a British collection may need all three. The published description can identify a sheet through its imagery. The acquisition file can identify a seller and sometimes a group price. The physical object retains evidence about how it was made or used that neither text exhausts. Combining them illuminates a specific transfer without forcing every document to answer the financial question.
Funds, donors and the people who owned the art

The words on Cleveland's pair of screens name the Severance and Greta Millikin Purchase Fund. They establish an acquisition by purchase associated with that fund. They do not establish that the Millikins owned the screens immediately before the museum acquired them. Nor do they give the amount withdrawn from the fund. Looking closely at the two records, the suffixes .1 and .2 separate the right screen from the left. The fund name is repeated across them. It would be unsafe to infer two independent payments merely from the fact that a collection management system displays two records with identical credit information.
A Japanese sword guard in the Metropolitan Museum provides another instructive credit. Accession 36.120.12 entered with the Howard Mansfield Collection as Mansfield's gift in 1936. Someone researching an earlier purchase by Mansfield would need papers documenting that transaction. The credit itself is already quite precise about a later one. It names a donor and the manner in which the object reached the museum. A curator could describe the history of collecting through such a guard without pretending to know the missing figure. Indeed, a precise acknowledgement of what the accession establishes makes the search for dealers' invoices or correspondence much more purposeful.
Consider how an account for a guard might look if a future researcher did find the dealer's papers. The dealer might have sold several fittings together, accepted a trade, quoted a price before allowing a discount or issued a receipt after an earlier deposit. Each possibility suggests a question to ask of the documents. None is evidence about the Mansfield guard by itself. The Freer papers have the unusual merit of letting us see the difference on the page. Price, commission, money sent to an intermediary, gifts alongside purchases and charges for transmitting money all survive in identifiable places. They make better guides to inquiry than any guess based on the modern appearance of metalwork.
A price that travels with its explanation
If there were room on a museum label for only one sentence about the Freer pair, the most informative sentence would name the buyer, the pair, the base amount, the commission and the year. An expanded catalogue entry could identify the intermediary and distinguish the Japanese summer purchase from the Detroit paperwork completed in December. The dollar entry belongs there too, provided it is described as an amount entered in the account. No modern equivalent is necessary to make the historical event intelligible. The buyer's choice, the work of the intermediary and the accounting for two screens as a pair already tell a remarkable amount about what was being transacted.
One could ask a further question about purchasing power. Different comparisons would illuminate different things. An index of household prices would answer a question about consumption; museum acquisition accounts might answer a question about relative institutional commitments; a wage series would frame a question about earned income. Choosing one requires evidence for the place, currency and period as well as a clear statement of what the comparison is meant to show. The three-thousand-pound group purchase in London and the eight-thousand-yen pair in Japan cannot be made equivalent by setting the largest printed numbers side by side.
What survives most vividly from the 1909 papers may be the small variations in what the office needed to know. A clerk marked a payment received. A descriptive entry preserved colours on two screens and an attribution circulating in Japan. A shorter line gave the artist as unknown. A list of ceramics distinguished three neck shapes and indicated a gift. These observations could coexist because the papers had different purposes. Read together, they give more than a price. They tell us which price belonged to which action, who undertook it and how the works of art entered an institutional history that their makers could never have foreseen.
It is possible to cite a price succinctly while giving someone else enough information to challenge the interpretation. Include the archival series, the year, the individual sheet and the words identifying the object group. Add which column supplied the amount when more than one currency appears. A reader looking up Freer's pair can then discover that one entry describes the seller's identity and another supplies details of the depicted flowers. If a transcription is improved later, the trail remains usable. This is especially important for collections studied across languages, where a proper name, an accession prefix or the first words of an object description can change between editions.
Questions about purchase records
Who bought the Japanese floral screens in the 1909 voucher?
Charles Lang Freer purchased the pair through Samurai Shokai in Yokohama. The paperwork now belongs to the Smithsonian archives; the original transaction was a purchase by the collector.
Why are both 8,000 and 8,800 yen recorded?
The summary gives 8,000 yen for the pair plus ten percent commission. The resulting 8,800 yen and a dollar entry of 4,470 appear in the account.
Was the December bank transfer the price of the screens?
The bank recorded a larger remittance of 48,040.50 yen to Samurai Shokai and acknowledged a dollar amount including the cost of transmission. The object-level entry identifies a separate amount for the pair.
How did three Owari bottles enter the same account?
The page identifies the three vessels as gifts from Mr. T. Hara. Their inclusion among the records of the trip provides no individual purchase price.
What did the British Museum pay for Anderson material?
The 1881 purchase of a group of works from William Anderson was £3,000. A separate acquisition of 311 illustrated books in 1882 cost £360, according to the scholarship cited below.
Are the illustrated Cleveland screens the same pair that Freer acquired?
The illustrated work is a different pair by Watanabe Shikō, Cleveland accessions 2021.132.1 and 2021.132.2. Its two records share a purchase-fund credit.
Museum films
Smithsonian National Museum of Asian Art — Freer’s Asian Ceramics Reconsidered
Smithsonian National Museum of Asian Art — Two Great American Collectors of Chinese Ceramics: Morgan and Freer
Collection records and further reading
- Smithsonian: Freer papers, summary for purchases and gifts
- Smithsonian: description of the floral screens
- Smithsonian: Peoples State Bank transfer to Yokohama
- National Museum of Asian Art: Accounting for Art
- National Diet Library newsletter: William Anderson and collecting Japanese art
- Michelle Huang: research citing the British Museum Central Archive
- Miya Itabashi: the 1881 and 1882 acquisitions
- What the 1891 Ives and Walters Gift Brought to New York
