Open the sword volume of Shūko jisshu at digital frame fifteen. Three long objects occupy one leaf, two the other. Short vertical annotations sit beside their upper ends. Now open volume four of a differently arranged edition at frame eighty-one. There they are again, in the same sequence and with writing in the same places. Even the little projections at the sides of the mounted forms seem familiar. It takes only a moment to recognize the design. Deciding whether a printer took both images from the same piece of wood is a much more interesting problem.
The Japanese National Diet Library preserves both books. Ikubunsha published the first in a series indexed under 1905. Kokusho Kankōkai published the second in 1908. Our four illustrations reproduce two pairs of openings from these separate library objects. Readers can move between the full images and check the arrangement without trusting a modern editor’s redrawing. The question at stake is practical. When a historian finds the same named sword in two printed catalogues, has a second observer confirmed the shape, or has a single visual witness acquired a second printing history?
Five forms that travel together

The first opening permits a simple count. Three images stand inside the left border and two inside the right. Their upper ends differ, and the little columns of writing do not all start on a common line. The eye can move from a dark band on one of the forms to a projection on its neighbour. This is information about the layout as well as the individual objects. A scholar drawing the same five items independently could preserve their shapes while choosing quite different intervals between them or an entirely different place for the explanatory writing.

The second opening preserves much more than the subject. The distribution between the pages, the mixture of writing and blank space, and the left-to-right relationships within each leaf correspond closely. These are grounds for recognizing a transmitted composition. The surrounding pages, however, immediately tell us that we are holding a different publication. A scan of the later volume carries this opening deep into a larger gathering of material. The earlier volume announces its subject as swords. The later one places swords alongside other categories. A page design has remained recognizable while the editorial setting has changed.
Notice the two different borders. The rules printed around the image belong to the composition of the book. The dark vertical interruption at the gutter belongs to the bound opening as it was photographed. In some places the image area appears very close to the fold. Elsewhere a strip of paper or a shadow runs between them. If two digital frames are placed one on top of the other and aligned only by their gutters, a false difference can run through the entire comparison. Printed marks and the geometry of a bound volume must first be separated with the eye.
The caption can be more useful than the largest shape. A single long curve can be copied quite faithfully by someone interested in the object’s overall form. An eccentric gap between characters, a tiny turn at the end of a stroke or a caption that begins at an unexpected height is less likely to recur by independent observation. A whole cluster of such coincidences strengthens the case for dependence. It still leaves open how the dependence was achieved. Copying a printed page by photographic means could preserve several of these accidents without any old block entering the later workshop.
What carries over when a block changes hands?
A woodblock preserves a surface from which a printer can produce another impression. The printed page preserves the result of a particular encounter between that surface, ink and paper. The distinction is basic but consequential. Ownership of an impression does not imply ownership of a block. A publisher able to issue a new book might have acquired blocks, arranged access to them or commissioned a reproduction from a copy of an earlier book. A modern library record identifying that publisher establishes responsibility for the edition; it does not serve as an inventory of the publisher’s equipment.
The Art Institute of Chicago’s study of Japanese illustrated books discusses colophons that identify a publisher or the owner of the blocks. Such information can survive at the end of a volume whose pictured leaves look virtually identical to those of another issue. It gives the historian a second line of enquiry beyond inspection of the design. Who represented themselves as entitled to publish? Was the statement particular to this edition, carried over with an older page, or supplied on a replacement leaf? The surviving copy and the wording of its colophon are both needed to answer.
One can see why the difference mattered to a reader interested in swords. An illustrated record could put a rare object before people who could never inspect it in person. The image acquired authority through repeated use, particularly if the caption named an owner or a place of preservation. A later volume might carry the same owner’s name long after its accuracy had ceased to be certain. The repetition is valuable testimony about what the publisher transmitted. Unless someone rechecked the object, it supplies little evidence about its location in the year the later book appeared.
Another pair, with more places to compare

Frame twenty-five in the Ikubunsha volume leaves the ranks of long mounted profiles behind. Two elongated rectangular forms appear alongside a vertical group of lobed shapes. Across the gutter are several smaller objects, with short entries occupying some of the intervening space. The visual rhythm is quite different from that of the five tall forms. The page can therefore test whether a second publication borrowed just one attractive plate or transmitted a more extensive sequence. A single repeated image is interesting; two quite different repeated openings a number of leaves apart tell us appreciably more.

In frame ninety-one of the 1908 publication the same two-page division and grouping reappear. Compare the central openings in the lobed forms, then the small outlines and the printed characters beneath the lower shapes on the facing page. All can be located without knowing the names of the represented objects. That is an advantage. A historian who starts by finding a famous name can overlook a displaced caption or identify a similar fitting too eagerly. Here the objects, their neighbours and the shape of the empty paper supply independent points for checking the match.
The difference between frame fifteen and frame twenty-five in the first book is ten. In the second book the corresponding frames, eighty-one and ninety-one, are also separated by ten. This small numerical observation is suggestive because the visual matches independently establish the endpoints. It is not proof that every intervening leaf is identical. A replacement, insertion and omission could in principle leave the same total distance. Nor do the frame numbers give historical plate numbers. They identify digital exposures of these particular bound copies, including anything photographed before the illustrated pages.
A catalogue entry for one of the small forms could now cite the two witnesses in a way that adds historical information. It might state that the arrangement appears in the sword volume issued by Ikubunsha and recurs in the larger Kokusho Kankōkai volume. It could identify both openings and distinguish the volume dates. If it instead described the two pictures as separate portraits agreeing on an object’s appearance in two different years, it would discard the most interesting inference the pages make possible. The likeness is evidence about the image before it is evidence about the metal.
The diagnostic value of a broken line
A technical study by Capucine Korenberg at the British Museum supplies a good independent comparison. The subject is Hokusai’s Great Wave, not a sword book. The research examines multiple impressions and asks which recurring changes can sustain an argument about the sequence of printing. One earlier investigator had recorded nineteen successive breaks around the title enclosure. The later study found only three breaks that appeared systematically in the impressions examined. The disagreement matters for anyone tempted to turn a faint spot on a single facsimile into a date for an entire edition.
The advantage of a population of impressions is that it lets one distinguish a recurring interruption from something that happened to one sheet. A printed line can look faint because too little ink reached one portion of the paper. A later photograph can lose the lightest traces. A damaged sheet can remove a small area after printing. In a series of independent impressions a break that returns in the same position and in a coherent sequence becomes a different kind of evidence. Even then a replacement or repair to a block may complicate a simple progression from sharp to worn.
Our two pairs are a much smaller body of evidence. The scans show differences in the apparent strength of the borders and the density of dark areas. Some marks near a binding are harder to follow in one image than in the other. Differences of this kind tell us where closer examination could be productive. They do not establish an order of wear between the two editions. The date order is already securely supplied by the books. It would be circular to assume that the later page must look more worn and then select faint marks to prove it.
The most promising additional witness would answer a specific unresolved question. A third impression of the earlier issue might show whether a supposed gap was already present repeatedly. A later issue with an explicit statement about the disposition of the blocks could give a documentary context for the match. A surviving block, securely identified and comparable in its local features, would offer evidence on an entirely different physical surface. More pictures alone are helpful only if we know what their relationship is to the sheets already under comparison.
What waste paper can say
A National Diet Library investigation approached Shūko jisshu from a less obvious direction. Its subject was printed waste reused inside other bindings. Material on the inside of a cover can preserve fragments that once belonged to a quite different publishing project. The library compared surviving portions of designs with illustrations from the antiquarian survey and considered the associations of seals and former owners. Such a study can recover evidence about the circulation of printed matter in places where nobody thought they were keeping a catalogue of swords or antiquities.
The object of comparison is important here. A fragment behind a cover may preserve the edge of an image or a partial caption. A matching image in a complete book can identify the printed design and the region from which the fragment came. The reuse subsequently visible in the binding is reuse of a sheet of paper. It is an entirely different event from making a second impression with the same woodblock. The two histories can intersect, but a piece of printed waste could survive even if the block had been destroyed immediately after the first printing.
The investigation also identifies a limitation of monochrome images made from earlier microfilm. Certain parts of a binding were not photographed in some sequences, and differences of colour cannot be read from a monochrome reproduction in the way they could from the original. This has a curious consequence for searching. A surviving cover lining can exist on a shelf, be crucial to the identification of a fragment and still be effectively absent from a digital comparison because the photographer never opened the cover in the required way. Absence from the viewer does not settle absence from the book.
The changing work of the publisher
A sword catalogue could be a specialist book, one part of a survey of antiquities, an exhibition handlist or a new selection designed for readers who knew the names but had never seen the objects. The role of an old image changes in each setting. In a specialist sequence an adjacent page can invite comparison of similar objects. In a general survey, a much wider classification and a new table of contents can make the same picture evidence for the range of material culture. A later publisher can produce a new intellectual arrangement while leaving a borrowed image almost untouched.
The Smithsonian’s study of Japanese book publishing provides a wider context for this distinction. An illustrated book was made by people performing different tasks, and its survival depends on the history of a physical copy. An edition can be represented today by a volume with a replaced cover, a separated part or an owner’s mark never present in another example. Students encountering a digitized copy through a page link can easily lose that sense of a book assembled, used and altered over time. Referring to the specific library witness brings some of that history back.
There is a useful editorial test for anyone writing about a repeated plate. Read a caption aloud with its nouns made explicit. Does the date belong to the object, the original drawing, the printing of the volume or the photography of the copy? Does “the same” refer to an object, an arrangement of designs, a printed image or a block of wood? A sentence that sounded precise when those nouns were implicit may prove to rest on an identification nobody has made. Giving the dates and the page references usually produces better prose as well as a more accurate claim.
A conclusion that can survive another impression
For the paired openings shown here, the affirmative finding is strong. Two separately catalogued editions transmit the same five-object arrangement and, ten frames farther on, another densely annotated group. The second publication preserves the reader’s path through these two points in the sequence while assigning the material a new position inside a larger book. A bibliographer can describe that relationship without settling the manufacturing history of every leaf. It is substantive evidence of how an image tradition travelled between publications, and of what a reader in 1908 could see again.
The physical identity of the printing blocks remains open on this evidence. That is no failure of the comparison. It distinguishes a claim that can already be demonstrated from one for which a different witness is needed. If a future researcher discovers a publisher’s account of acquiring the blocks or a telling pattern shared among several physical copies, the argument can advance without undoing anything established here. The full openings, the frame references and the names of the two volumes make the result available for that next reader to test.
One final detail is worth carrying away. The space between two images can travel as faithfully as either image itself. So can the slightly awkward placement of a short inscription beside a curved outline. Those modest features rarely attract the eye of someone looking for an impressive historical blade. They are exactly what a historian of books needs. They show that what survived across publications was a way of laying out knowledge on a page, an invitation to compare that can outlast the first readers, the publishers and possibly the very blocks from which they printed.
There is a small practical reward for recording the second pair as well as the first. A reader whose copy has lost its opening leaves may still recognize the group of lobed fittings farther into the book. The combination of pictures and nearby writing provides a route to identification even where the original title slip is missing. It is precisely in incomplete copies, which can seem the least promising witnesses at first, that a carefully described pair of openings may ultimately prove most useful.
Questions about blocks and editions
Which two library books are compared?
The Ikubunsha sword volume of Shūko jisshu is compared with volume four of the 1908 Kokusho Kankōkai publication. Each has its own National Diet Library record.
Which openings show corresponding designs?
Ikubunsha frames 15 and 25 correspond to Kokusho Kankōkai frames 81 and 91. The digits refer to frames in the digitized copies.
Do the matches prove that the physical blocks were the same?
They establish transmission of the pictured design and arrangement. A copy made from an earlier printed image can also preserve those relationships.
Why compare the positions of short captions?
Writing, spacing and minor irregularities can distinguish the transmission of a whole page from two independent drawings of similar objects.
What can printed waste inside a binding reveal?
A fragment may match an identifiable design in a complete publication. Reuse of the printed sheet is distinct from reuse of its printing block.
Why can a pale line fail to establish a later printing?
Inking, paper, damage to a sheet and reproduction can all affect its appearance. A recurring pattern across independent impressions is stronger evidence.
Museum films
The British Museum — How to print bokashi | A meditative journey into Japanese woodblock printing
The Met — Sunday at the Met: Brush Writing in the Arts of Japan
Collection records and further reading
- National Diet Library, Ikubunsha sword volume
- National Diet Library, Kokusho Kankōkai volume four, 1908
- National Diet Library Monthly Bulletin, September–October 2025, printed waste in bindings
- Capucine Korenberg, comparison of Great Wave impressions
- Art Institute of Chicago, Conservation at the Art Institute of Chicago
- Smithsonian, illustrated books and publishing
- Drawing the Curve: How Historical Catalogues Illustrated Sori
