I named my software Soumille after the author of the most useful and accessible book about the game. In Le Grand Trictrac, Soumille teaches trictrac through the device of games played between the characters Damon and Cloris. There are some 270 woodblock prints illustrating the games move-by-move, making it quite easy to learn the rules and, to some extent, strategy. For those interested, I have published English language rules here, and reproduced Soumille's first game, here.
Le Grand Trictrac was first published in Avignon in 1738; further Avignon imprints appeared in 1739 and 1756. The book was published in Paris in 1756, 1766, 1790, and 1801 (the links are to full-text copies available for download from Google). I have several copies of this book including the Avignon edition of 1739. I had always wondered why a second Avignon version appeared just one year after the first. It seemed unlikely that the book had sold out and needed to be reprinted. And my copy of the 1739 version had five cancelled leaves, suggesting that it was a reissue of the 1738 edition with a new title page and other changes.Why did the book reappear so quickly?
Until this week, I had never seen the first edition of 1738, but I was able to buy a copy at the spectacular auction of the Messager collection of gaming literature and equipment from the auction house Alde in Paris. When it arrived, after making sure it was complete, I began to compare it with the my 1739 version. With the two books side by side, it quickly became clear that they are the same setting of type, and hence by definition both first editions, but different issues.
The first issue, pictured at right, has an unsophisticated but charming frontispiece of Damon and Cloris playing trictrac.
The second issue lacks the frontispiece and other preliminary material, but from page one of the text, the type is the same as that of 1738 except for the cancels.
I have written about cancels many times--they are great fun for bibliographers. To correct some sort of problem, the printer would print a new page; the binder would cut out the incorrect leaf, leaving a stub, and paste the replacement on the stub. As I wrote in the essay "Every Cancel Tells A Story, Don't It (part 1)", what I find most interesting is not the fact of the cancel, but the reason for it. With both the 1738 and 1739 issues of Le Grand Trictrac, I could compare the original and replacement pages side by side and learn the reasons for the cancels.
In the first cancel below, the woodblock of the trictrac board is wrong in the page on the left from 1738. Note the upper right quadrant where two black checkers belong on the point marked "H" but are actually shown one point to the right. The diagram does not match the textual description of the game. On the right is the corrected diagram of 1739. If you look closely (as always, you can click to enlarge it) at the 1739 page, you will be able to see the stub from the excised leaf. A close look will also show that the type has been reset for the page.
|leaf E2v page 36|
|leaf Ll1r page 265|
Another incorrect diagram appears on page 265...
|leaf Oo2v page 292|
...and a third on page 292, where again the stub is prominent.
I'll skip one cancel where there was a minor textual change, but my favorite cancel is this:
|leaf Y4v page 176|
The diagram is unchanged. Well mostly unchanged. More about that in a moment. What is new is the addition of text below the diagram. Cloris rolls 22 and scores 4 points for being able to "hit" Damon's blot on the point marked P with a checker on the point marked L (see the rules describing "battre sur une demi-case"). Not mentioned is that Cloris can also score six points for "battre le coin" and that omission must have vexed an attentive reader.
When Cloris fails to score the points, as she often did in the sample games, Damon could claim them, scoring "écoles" or sending Cloris to school. But no mention is made of écoles here. In the cancel, neither player could be given the points without changing the flow of the game, so Soumille wrote that Cloris hit the "coin" but didn't notice, and neither did Damon! Of course it was Soumille who at first didn't notice. An amusing omission and an unusual reason for a cancel.
There is one more point to notice about these side-by-side pages with the unchanged diagram. Notice that the dice have moved although they display the same roll. And the figure number seems to have drifted off to the right. That suggests that the piece of wood used to print the diagram did not contain the dice or the figure number. The dice were likely separate pieces of wood placed on the diagram and the figure numbers were pieces of type place in the woodblock.
I've looked quickly at later editions. The 1756 Avignon edition had new, smaller woodblocks. while the Pairs editions of 1756 and 1766 had a third set of woodblocks. What a lot of work for the woodcarver!
I've often noted that Hoyle wrote the first book of strategy on any card game, whist in 1742. He also wrote the first book of backgammon strategy in 1743. But, as we see from Soumille, there was an instructional book on trictrac four years earlier. And like the early Hoyles, the early versions of Soumille reveal much about their publishing history.