TNS news nishikie

Cherubs and seals

By William Wetherall

First posted 1 September 2005
Last updated 1 February 2010

Correlating cherubs and seals    Pattern analysis, initial observations, problem, method, data
Tables    Cherub style and seal type distribution by issue number  |  Cherub style and seal type distribution by year and month

Correlating cherubs and seals

Pattern analysis

All elements in a body of work lend themselves to analysis. Stylistic features, pigments, even themes usually change over time, in ways that reflect the conditions under which the work was produced. Typically such changes are patterned, and the patterns can be seen as stages in the growth or decline of the work.

One can choose any elements, and any number of elements could be taken together. However, at least two elements are needed in order to determine how change in one might be reflected in, or influenced by, change in another. And since the baseline of change is time, at least one of the elements chosen for analysis should be a date or at least be datable.

Analyzing TNS nishikie

Series of works like Tokyo nichinichi shinbun (TNS) and Yubin hochi shinbun (YHS) news nishikie are ripe for such analysis, for many issues of each series were produced in the course of its life by the same drawer and publisher, and one has to expect to find changes in design features as the series developed. Here we will focus on the TNS series.

The TNS series undergoes four stages of change defined by three types of cherubs (see "TNS Four Stages" for details). This is obvious just by looking at the prints in the order in which they are listed in tables and databases (Kanbara, Ono, Nishigaki, Tsuchiya), namely by the number on the banner. The problem is that the change does not strictly correlate with the banner number, because it is only the issue number of the TNS broadsheet, which came out daily, and does not signify the order of publication of the TNS nishikie, which marched to its own drummer.

Initial observations

Our analysis begins by observing that, though there are four stages of banner/cherub design, the distribution of cherub styles by issue number breaks down into the following five groups.

First TNS series

Group 0    0000 (Unnumbered)   No banner, big cherubs
Group 1    1-220   Stage 1 cherubs, Typical graphs    Stage 1 banners
Group 2    322-781   Stage 1 and 2 cherubs, Typical graphs    Stage 1 and 2 banners
Group 3    813-1036   Stage 2 cherubs, Typical graphs    Stage 2 banners
Group 4    9001-9002 (Unnumbered)   Stage 2 cherubs, Typical graphs    Stage 2 banners
Group 5    1043-1060   Stage 3 cherubs, New bold graphs    Stage 3 banners

Second TNS series

Group 6    9002-9006 (Unnumbered)   Stage 2 cherubs, Different graphs    Stage 4 banners

I have omitted TNS-0, which is exceptional in its design. I have also ommitted variations of prints (such as those of TNS-851 and TNS-9003) since the print particulars of the variations are the same. The four stages in the life cycle of the TNS nishikie series have been discussed elsewhere (see "TNS Four Stages"). Here we want to focus on resolving the apparent confusion of Stage 1 and 2 cherubs in Group 2.


The pressing question is whether the apparent overlap of Stage 1 and Stage 2 cherubs in Group 2 is the result of fickle use of the two cherub designs during an experimental period when the two designs were rivals, so to speak -- or whether we can discover objective criteria for dividing Group 2 prints into two sub groups, one published earlier with Stage 1 cherubs, and the other published later with Stage 2 cherubs -- thus resolving the confusion.

The confusion, of course, might be real. Perhaps there was a period when Stage 1 and Stage 2 overlapped, when both cherub designs were used. But this does not accord with what we know about how publishing houses ordinarily operate. When the design of a publication series is changed, the new design is usually introduced in lieu of the old design, which is no longer used, except when reprinting older blocks or plates.

We are also motivated to resolve the apparent confusion in Group 2 prints because, if the style of the banner on a given TNS print was not a matter of weather or mood, but reflected a decision to use a new design from a certain day, then the style of cherub on a print might be helpful in dating prints that don't have an approval seal. And herein lies the ultimate object of our analysis.


Obviously we cannot use the issue number to divide Group 2 into earlier and later prints. But what about the approval seal? Is there a patterned relationship between cherub styles and approval dates? If so, the pattern should be evident in a table showing the two variables -- cherubs and seals -- side by side.

Instead of annual growth rings or Carbon 14, TNS prints have aratamein (approval seals) and a few have "otodoke" () or "reported" dates. Only the last three known TNS prints reflect revisions made on 3 September 1875 to a 1869 regulation that governed printed matter. The revisions replaced the aratamein system with the otodoke practice of registering planned publications with the government. All prints in Group 2, however, were published well before this revision. So any clues they yield as to when they might have been published, other than the issue number (at best a "no earlier than" index), would have to be in their approval seal and/or cherub design.

One wrinkle in the use of seals to date a print is that there may be more than one seal design for a given month. The key to our analysis turns out to be variations in the seals used for the 10th month of 1874. This was an "inudoshi" (N) or year of the dog (). Seals for this year combined one or another style of "inu" () with a number for the month. Some 1874-10 seals used the Hindu and Buddhist "manji" () to denote "ju" (\) or "10", while others used a "rotated ju" (w). For an overview of approval seals and a look at the seals in question, see Seal of Approval and Date of Print in the article on "Typical Print Details".

As we shall see, not only does cherub design seem to correlate with date, but cherub design during the 10th month appears to correlate with seal design. In fact, the "dog manji" () seal appears to have been used only on prints with Stage 1 cherubs, while the cline marking the transition from Stage 1 to Stage 2 cherubs seems to fall later in the month when the "dog ju" (w) seal was used.

This distinction is not as trivial as it may seem. In fact, it will help us rank certain prints -- namely those related to the Taiwan expedition -- in a way that makes more sense when trying to understand when they were published in relation to one another and the events they depict. As we will see below, the significance of the order in which prints were published is quite different from the meaning of the order in which the original stories appeared in the broadsheet (as shown by the issue number on the banner of the nishikie).


The following table shows a breakdown of TNS-1 thru TNS-856b by (1) Issue number, (2) Cherub style (stage), (3) Seal type (shown only for prints with approval seal for the 10th month of 1874, in which "10" was represented as either "" or "w" in the seal design), (4) Seal date (year-mo), (5) Lunar story date (year-mo-dy), and (6) Solar story date (year-mo-dy). Lunar dates are shown only until the 3rd day of the 12th month of 1874, which was declared the 1st day of the 1st month of 1873 on the solar calendar, from which time all dates are solar.

The data in the table was compiled by examining all the prints in Tsuchiya Reiko's Nihon nishikie shinbun shusei (Tokyo: Bunsei Shoin, 2000), a CD-ROM which contains large images of practically all known news nishikie. The CD-ROM cross-indexes all prints by seal dates, and I was able to confirm that the list of TNS prints with 1874-10 seas is both complete and correct. However, the list does not break down 1874-10 seals into types, as I have here.

The lunar dates are from Ono 1972, as are the solar dates after 1872. The solar dates corresponding to the lunar dates are my own conversions. More such dates will be provided as they are found and confirmed.

I confirmed the seal dates using Ishii Kendo's 1920 guide, Nishikie no aratamein no kosho [Studies of nishikie approval seals], Tokyo: Unsodo, 1920, revised 1932. Ishii's study continues to be available in a facsimile edition of its original Japanese-style printing and binding, as it is still the bible for reading and interpreting approval seals. See Seal of Approval and Date of Print in the article on "Typical Print Details" on this website for a scan of the page in Ishii's book which shows the seals used during the years the TNS and YHS new nishikie were being published.


TNS cherub style and seal type distribution by issue number
Group 0 TNS 0 Unnumbered, no banner, big cherubs

This print, a promotional flyer, announced the start of the TNS nishikie series (Groups 1-5).

Group 1 TNS 1-220 have Stage 1 banners (Stage 1 cherubs, Typical graphs)
Issue Cherub
story date
story date
1 1 1874-10 M05-02-21 1872-03-29
3 1 1874-10 [1872-02] [1872-03/04]
40 1 1874-09 M04-04-04 1872-5-10
50 [48] 1 1874-09 M05-04-13 1872-5-19
101 1 1874-09 [1872-06] [1872-07/08]
111 1 1874-10 [1872-06] [1872-07/08]
185 1 [1872-09] [1872-10/11]
220 1 1874-08 M05-10-26 1872-11-26
Group 2 TNS 322-781 have Stage 1 and 2 banners (Stage 1 and Stage 2 cherubs, Typical graphs)
322 2 Banner w 1874-10 1873-03-22
428 2 Banner w 1874-10 1873-07-15
431 1 1873-07-23
445 1 1874-09 1873-08-07
472 1 1874-08 [1873-09]
491 2 Banner w 1874-10 1873-09-28
512 1 1874-10 1873-10-22
566 2 [1873-12]
592 1 1874-01-26
656 2 1874-04-09
687 2 w 1874-10 1874-05-14
689 2 w 1874-10 1874-05-16
694 1 1874-05-22 
697 2 1874-05-25
708 2 1874-12? [1874-06]
712 2 Banner w 1874-10 Taiwan Expedition
Battle of Stonegate
723 1 1874-10 [1874-06]
726 1 1874-10 Taiwan Expedition
Botan girl
736 1 1874-09 Taiwan Expedition
Kishida Ginko
742 1 1874-10  [1874-07]
748 1 Banner w 1874-10 [1874-07]
752 1 w 1874-10 Taiwan Expedition
Natives kowtow
754 1 1874-07-25
781 1 1874-09 1874-08-26
Group 3 TNS 813-1036 have Stage 2 banners (Stage 2 cherubs, Typical graphs)
813 2 Banner w 1874-10 1874-10-03
822 2 Banner w 1874-10 1874-10-12
824 2 1874-11 [1874-10]
832 2 1874-10-23
833 2 Banner w 1874-10 1874-10-24
838 2 1874-11 1874-10-29
847 2 1874-11 Taiwan Expedition
Showing the flag
849 2 Taiwan Expedition
Peace negotiations
851 2 Taiwan Expedition
Soldier's ghost
856a 2 1874-11-19
856b [858] 2 1874-12 1874-11-22

TNS 860-1036 are omitted for purposes of this analysis.

Group 4 TNS 9001-9002 (Unnumbered) have Stage 2 banners (Stage 2 cherubs, Typical graphs)
TNS-9001 2 [187?-09-13]
TNS-9002 2 w 1874-10 Taiwan Expedition
Trophy heads
Group 5 TNS 1043 to 1060 have Stage 3 banners (Stage 3 cherubs, New bold graphs)

TNS 1043-1060 are omitted for purposes of this analysis.

Group 6 TNS 9003-9005 (Unnumbered) have Stage 4 banners (Stage 2 cherubs, Different graphs)

TNS 9003-9006 are omitted for purposes of this analysis.
These prints constitute a second, different, and later TNS series.
9006 has a cartouche rather than a banner and has no cherubs.


TNS cherub style and seal type distribution by issue number
Seal and notification dates by year and month
None 1874 1875 1876
Stage Issues Prints None 8 9 10 10w (b10w) 11 12 1-8 11-12
Stage 1 banners (Stage 1 cherubs, Typical graphs)
1.1 1-220 8 1 1 3 3
1.2 322-781 14 4 1 3 4 2 (1)
Stage 1 totals 22 5 2 6 7 2 (1)
Stage 2 banners (Stage 2 cherubs, Typical graphs)
2.1 322-781 10 3 6 (4) 1
2.2 813-895 22 11 3 (3) 3 5
2.3 900-1036 42 20 22
2.4 9001-9002 2 1 1
Stage 2 totals 76 35 10 (7) 3 6 22
Stage 3 banners (Stage 3 cherubs, New bold graphs)
1043-1060 10 6 4
Stage 4 banners (Stage 2 cherubs, Different graphs)
4.1 9003 1 V on banner with Stage 2 cherubs 1
4.2 9004-9005 2 XV on banner with Stage 2 cherubs 2
4.3 9006 1 V cartouche (no banner, no cherubs) 1
Stage 4 totals 4 Unnumbered triptychs in Second TNS series (late 1876) 4
Totals and percents (TNS-1 to TNS-9006)
Totals 112 46 2 6 7 12 (8) 3 6 26 4
Percents 100 41 2 5 6 11 (7) 3 5 23 4
Stage Issues Prints None 8 9 10 10w (b10w) 11 12 1-8 11-12
None 1874 1875 1876
Seal and notification dates by year and month



A number of Group 2 prints that have no seals. Of course, they might have been assigned a "rotated ju" (w) seal. However, the data strongly suggest that the "manji" () version of the seal preceded the "rotated ju" (w) version.

On all but two prints, Stage 1 cherubs are associated with the 1874-10 seal that uses "manji" () to represent "10", while Stage 2 cherubs coincide with the 1874-10 seal with a "rotated ju" (w). The two exceptions feature Stage 1 cherubs with "rotated ju" (w) seals.

This is at odds with the order in which variations of 1874-10 approval seals are shown in Ishii Kendo's 1920 guide to seals (see above). Ishii lists three seals for the 10th month of 1874: one "rotated ju" (w) seal, followed by two variations of a "manji" () seal. It is not clear that he intended their order in his list to represent their order of appearance and use.

The seal on TNS-708 is smudged. I have called it 1874-12 after Tsuchiya 2000 (Bunsei Shoin CD-ROM). There are "ju" (\) and "rotated ju" (w) variations of the 11th and 12th seals for 1874, but they are not objects of the present analysis.

If it can be assumed that one day Yoshiiku or someone else in the Gusokuya group decided to change the cherub on the left, and that there was no vacillation about the change in cherub design, then the transition came during October 1874, after the "manji" () seals were no longer being used, and midway during the period when the "rotated ju" (w) seal was in use. If so, then the two prints which show Stage 1 cherubs with "rotated ju" (w) seals were published before the prints featuring Stage 2 cherubs with "rotated ju" (w) seals.

For an example of how the above analysis can be applied to improve conjectures about the publishing order of some prints, see TNS Taiwan Expedition prints.