Tuesday, July 14, 2020

Whist and its Masters

updated August 3, 2020 with information from the second Foster letter

This meandering essay will discuss an 18c article and a series of 19c articles on whist strategy. The earlier article has been quoted without attribution so many times, that I want to acknowledge the original source. The later series is, I think, unknown to today's scholars of whist.
Let's begin with an entry in Jessel's 1905 bibliography1 of gaming books:
560. FOSTER, ROBERT FREDERIC. - Whist and its Masters. In The Monthly Illustrator, Sept. 1896 to March 1897, inclusive. I. The Old School. II. The New School. III. The Signalling School. IV. The Scientific School. V. The Number-Showing School. VI. The Duplicate School.VII. The Private Convention School. (Butler, p. 43. These papers are about to be reprinted in book form.)
Let's unpack the entry a bit. Whenever Jessel had not seen a book he included in his bibliography, he was careful to note how he learned of it. Here, the parenthetical item "Butler, p. 43" means that he saw the book listed in William Butler's outstanding work, The Whist Reference Book (1898)2. The book is arranged as an encyclopedia and on pages 42-3 is an article called "Articles on Whist". The entry of interest is:
"Whist and its Masters," by R. Frederick Foster, Monthly Illustrator, Sept. 1896 to March 1897, inclusive. I. The Old School. II. The New School. III. The Signalling School. IV. The Scientific School. V. The Number-Showing School. VI. The Duplicate School.VII. The Private Convention School. 
So Jessel did no more than copy Butler.

As much time as I spent with Jessel, I had never paid much attention to this entry. Certainly if the papers had been reprinted in book form, I would have run across them at one time or another. Why did Jessel think they were going to be reprinted?

I visited the Jessel collection at the Bodleian Libraries, University of Oxford in 2018, As I looked at a lot of books and a lot of Jessel's handwritten notes about his collection, I discovered that dozens of his books included letters from the author which Jessel had pasted in the front of their books. I looked at as many of these letters as I could find and got a sense of how Jessel collected, and how he put together his bibliography. The most frequent correspondent was Foster and the earliest letter I found was in an edition of Foster's Complete Hoyle (1897), shelf mark Jessel e.443.

Foster wrote the letter on November 16, 1903 apparently responding to a request from Jessel that he list all the books he had written on card games. Foster mentions 20 titles and their publishers, but also describes "Whist and its Masters":
"Whist and Its Masters" was a series of articles, ten I think, published in a magazine called "Home & Country" I had arranged to have them appear in book form, but the publishers failed, and it fell through. These articles were a complete history of the strategy of the game, showing how it developed and enlarged, and the articles were illustrated by photos of all the men who had taken a leading part in advocating certain conventions. Some persons thought these articles the best things I ever wrote on Whist, I have no copies of them."
A later letter by Foster said that plan to publish the book was revived. The letter, dated January 1, 1905, is pasted in Foster's Practical Poker, shelf mark Jessel f.425. It is on the letterhead of The Sun, a New York daily newspaper to which Foster contributed columns on card games. "We [presumably The Sun] are going to reprint the 'Whist and Its Masters' and I shall be pleased to send you a copy for your library. How I sigh for that library!"

These letters intrigued me! I originally began to collect books on whist to understand the history of whist strategy. As I got more into Hoyle and its publishing history, that desire waned. There is not a lot of literature on the progression of whist strategy. William Pole's The Evolution of Whist (1897), here, does a creditable job, as does Butler's Whist Reference Book.

The revived plan must have failed as well--there is no hint that the book was ever published. So I needed to locate the original articles. I turned to ILL, that is inter-library loan, an experience both frustrating and rewarding in this instance. The frustration was that I did not have good bibliographical information about the articles. Butler and Jessel were wrong, the periodical is, as Foster indicated, Home and Country; its publisher is The Monthly Illustrator Publishing Co. of New York. The dates were wrong as well. The first article appeared in August 1896, the final article, number nine (not ten, as Foster recalled) appeared in April 1897.

ILL was able to provide digital copies of six of the nine articles. Surprisingly, I found a seventh on eBay for about $20. And, I have leads on the final two, but can't really pursue them until libraries reopen after the Covid-19 sheltering. Here is a complete list:
  • I. The Old School, 13:1, 15-21, August 1896
  • II. The New School 13:2, 89-94, September 1896
  • III. The Signalling School, 13:3, 153-7, October 1896
  • IV. The Scientific School, 13:4, 211-15, November 1896. 
  • V. The Number-Showing School, 13:5, 295-9, December 1896
  • VI. The Duplicate School, 13:6, 376-9, January 1897
  • VII. The Private Convention School, 14:1, 11-16, February 1897
  • VIII. The Common-Sense School, 14:2, 109-14, March 1897
  • IX The School of the Future, 14:3, 205-9, April 1897
I'm left to wonder who thought the articles were the best writing on whist Foster had done. Perhaps Foster himself?

Below is the cover for the issue I found on eBay. 

Home and Country
February, 1897

And here, a couple of pages from the Foster article, "The Private Convention School."

The articles are quite good, if one overlooks some overly-flowery prose. I'd like to focus on the first article as it is the one to focus on Hoyle and before. Before? Haven't I always claimed that Hoyle was the first to write on the strategy of card play? Foster writes:
The first attempts to reduce the practice of whist to a science appear to have been made by a coterie of players who met at the Crown Coffee-House, in Bedford Row, London, early in the last [18th] century, and of whom the first Viscount Folkestone is the best known. Unfortunately, they left no authentic record of the results of their investigations, and we have it on hearsay evidence only that they followed the general principles of "playing from the strongest suit (not the longest), studying the partner's hand, and playing to the score." (16)
The anecdote about Lord Folkestone and the Crown Coffee-House appears throughout the literature of whist. It is rare that any reference is supplied. Where did the story come from?

The source is an article by Daines Barrington in Archaeologia called "Observations on the Antiquity of Card-playing in England," a paper read before the Society of Antiquaries on February 23, 1786. Barrington traces the origin of playing cards and, at the end of the paper, he discusses the most popular card games: primero, ombre, quadrille, trumps, swabbers, and lastly whist.

Of whist, Barrington writes:
...[W]hisk seems never to have been played upon principles till about fifty years ago, when it was much studied by a set of gentlemen who frequented the Crown coffee-house in Bedford Row. (145)
A footnote begins below and continues on the next page:
I have this information from a gentleman who is now eighty-six years of age. The first lord Folkstone was another of this set. They laid down the following rules: To play from the strongest suit, to study your partner's hand as much as your own, never to force your partner unnecessarily, and to attend to the score. (145-6)

Barrington (145)

Barrington (146)

Two comments. First, I stand by my claim that Hoyle was the first to write about the strategy of whist. The Crown Coffee House group may have been the first to form principles of play, but they did not write or publish on the game. Second, one sees the suggestion that Hoyle may have been part of the group at the coffee house. That is certainly not true. Hoyle was a household name in 1786 when Barrington's piece was written. Had Hoyle been part of the group, Barrington's "gentleman friend" would have told Barrington; Barrington would have added Hoyle's name alongside that of Lord Folkstone.

I'm delighted to have tracked down (most of) the series of articles by Foster. They are well-written and informative. If you manage to locate a copy, be aware of a caution Butler made about Foster's writing in the Whist Reference Book:
[Foster] is also a frequent contributor to other publications, his recent series of articles (1896-'97) in the Monthly Illustrator, ...containing much valuable and interesting material, although tinctured with his likes and dislikes, which are very strong. (184)

1A Bibliography of Works in English on Playing Cards and Gaming. London: Longmans, Green, and Co. 1905. Available for download.  
2Available for download. If you happen to click through to the Google copy, be sure to check out the smile-inducing pages 16 and 17.

Monday, April 13, 2020

The First Piracy. Part 5. Bibliographical Concerns

updated June 29, 2020
This is the fifth and final essay on the first piracy of Hoyle's Whist. I have identified lots of variants in the surviving 44 copies. Here I ask the bibliographical questions--how many editions, issues, and states do these copies represent? The bible for answering these questions is Bowers.1 Indeed the mantra in my bibliography class was WWBD--What Would Bowers Do? His prose is dense, often argumentative, sometimes infuriating. Non-bibliographers who want a flavor of Bowers will be rewarded with frequent excerpts of his writing below.

To recap, part one provided an overview of the piracy and its physical structure. The second looked at gatherings B-H and K, noting that there were two settings of type for each of them, but one setting of I, L, M, A, and χ. The printer James Mechell printed two thirds of the book and decided to increase the print run. Most gatherings were reset; others were printed in larger quantity.

Essay three focused on gathering A, noting that even though there was a single setting of type, there were variations in the half title, the title, and the Advertisement. Essay four put things together. I was able to sequence the variants in gathering A, identify four phases of printing, identified by which variant of the title page is included, and summarize them in the chart below:

variations in gathering A

Variant 1 of the title page is the most common, appearing with variants 1 of the half title and Advertisement. For phase 2, the half title was reset, unlocking the forme and causing some type movement on title page, variant 2. That spacing was in part corrected in variant 3, when some type also slipped at the lower right margin of the Advertisement. Finally, the words "second edition" were added to produce variant 4 of the title page.

From examination or reports of 44 copies (recall that one copy lacked gathering A and cannot be included in the charts), I have identified which settings of the later gatherings combine with which phase of gathering A.

Setting of later gatherings grouped by title page variant

Before I delve deeply into Bowers' Principles, I should note that I'm not sure he would approve of the chart above. In a footnote, Bowers writes:
Attempts to link a press-altered title-page with certain press-altered formes of the text in another sheet usually betray bibliographical ignorance. (51)
I am linking a press-altered title page with other sheets that are reset, rather than altered in the press. As the data show, the correlation between the printing phases of gathering A and the resetting of the rest of the book are overwhelming. I trust that I am not betraying ignorance.

One of the jobs of a bibliographer is to classify copies of books by edition, issue, and state. Bowers defines these terms in a dense 88-page Chapter 2 of Principles. He notes that:
...books exist in separate editions and issues; parts of books may exist in variant states, although in certain special circumstances a copy of an edition or of an issue may itself be said to exist in a certain state. Bibliographies customarily give separate major headings only to editions, with issues listed under subheadings. Variant states are ordinarily treated under the heading or subheading to which they apply... (37)
Bowers defines the key terms:
An EDITION is the whole number of copies of a book printed at any time or times from substantially the same setting of type. (39)
An ISSUE is the whole number of copies of a form of an edition put on sale at any time or times as a consciously planned printed unit...(40).
...STATE is synonymous with VARIANT, and can be applied to any part of a book exhibiting variation in type-setting...(41)
How do these concepts apply to the first piracy of Hoyle's Whist? How many editions? Are there any separate issues? Which of the variants constitute a separate state? It's certainly not obvious.

Bowers does allow that no definition of "edition, issue, and state will indicate the invariable line to take with a small number of abnormally complex books..." (38) Perhaps this is one of those? Indeed trying to apply the definitions to the piracy has made me a little crazy. At various times over the past decade or so, I've come to different conclusions. Let me lay out my current thinking in this essay. I welcome comment from other bibliographers.

I am going to approach the problem in a series of steps. First, I will eliminate the outliers. Looking at the chart above, it is clear that almost all the copies with a variant 1 title page have first settings of the later gatherings. Almost all the copies with variants 2-4 have second settings. The outliers were introduced not when the book was printed, but when it was bound. After printing, the sheets must be dried, cut into half sheets, folded, and brought together for sewing. It is easy to see how the odd second setting might find it way into a book composed primarily of first settings. Eliminating the outliers simplifies the problem:

Eliminate outliers

Bowers does not address outliers introduced during binding specifically. He does examine examples where different editions sheets are bound indiscriminately (110-1), but not where the edition sheets, as here, are bound nearly uniformly. Bowers warns again letting differences in binding dictate new states (43-4). I am very comfortable ignoring the outliers noting them in the bibliography as copy-specific differences rather than as separate states.

My second step is to look at the differences between variants 2 and 3 of the title page A2r, which occur with differences in the Advertisement on A3r.



Perhaps the leftward movement of the T was deliberate. In the advertisement, the semicolon after "Undertaking" popped out of the forme and the hyphen and "l" from the next two lines migrated upward. As I mentioned in part 3, I have no idea where the comma after the catchword "'vai" came from. Bowers deals precisely with this situation:
Individual types were sometimes jerked from the forme by the ink-balls. Sometimes they went unnoticed...(47)
Such a change, Bowers notes, does not affect edition or issue, but state, so we can collapse variants 2 and 3 as we think about edition and issue:

Collapse variants 2 and 3

What about variant 1 versus variants 2-4? This is when Mechell decided to expand the print run, resetting gatherings that had already been printed and distributed. Bowers address this situation in his discussion of edition, but there is a circularity is his language that bothers me:
In quite a different category are books where the type was not left standing by design from the start, or from a point early in the printing, but had been normally undistributed when a decision was made to reset and print a new edition. (109)
It is the word "edition" that bothers me. The decision is to reset and print more copies. Whether they constitute a new edition is precisely the question. Bowers continues:
This decision may have been reached towards the end of the first printing and some formes kept standing by design, but the difference is immaterial. The characteristic of these books is that the standing type is from the last sheets to be printed (including preliminaries when these were last printed) [I, L, M, A, and χ in our case] and is not scattered throughout the book in a manner clearly showing that the lack of distribution was abnormal. (109)
He gives a number of 16c and 17c examples including one2 where the "title page was used in the identical setting in the second edition" along with some standing type. As I look at surrogates of that example, I see that what Bowers and other sources call the second edition, does have a title page identical with the first. This is exactly the Whist piracy--enough of the later sheets have been reset after variant 1so that it is not substantially the same setting of type. It is a second edition, even though one would not discern that from the title page.

Now what of Phase IV when the words "second edition" were added to the title page? Books from Phase IV are substantially the same setting of type as Phases II-III, so they cannot be a third edition. Could they be a separate issue?

Generally, Bowers wants a change in title page to create an issue. I'm going to ignore his distinction between separate issue and reissue, but will give you a flavor of his language:
We must, therefore, arbitrarily assume that any alteration made in the form of a book which was not important enough to justify a new title-leaf to call attention to it or to take advantage of the opportunity to bring the book up to date is a printer's attempt belated to construct an 'ideal copy' of his original issue and is not a reissue in which sheets are given new life or chronicle change in publishing conditions by alteration of form. (67)
But is the change in title page for phase 4, adding the words "second edition" to the same setting, enough to trigger another issue? Bowers would say no, although he doesn't have an example that matches exactly. He considers stop-press alterations of an imprint or date on the title page as creating a separate state, not issue:
Alteration of an imprint, usually for a change in a printer's, bookseller's, or publisher's name, or in the date, when performed by stopping the press comes in the same category as any stop-press alteration...If we are to regard the distinction between state and issue as necessarily made on bibliographical grounds whenever possible--that is, on the printing of the book rather than on less tangible considerations--alteration of a date or of a name in the imprint by stopping the press during the impression definitely comes under state* since it was demonstrably performed during continuous printing. (50)
*In a long footnote, Bowers notes the difference between bibliography and cataloguing, saying in part "...for cataloguing purposes variation of any kind on the title-page, if observed, constitutes a different issue. The use of the term issue in two quite different senses by cataloguers and bibliographers is undoubtedly of the greatest inconvenience for scholars and confusing to the student, but it is scarcely an argument that bibliography should limit itself to elementary cataloguing standards in order to procure uniformity." (50-1)
I am confident that Bowers would treat the stop-press insertion of "second edition" as a bibliographer, not a cataloguer; as a state, rather than an issue; .

Certainly the exposition of the story is more important than the bibliographical classification. Nonetheless, I do need to organize this section of the bibliography by the usual constructs of edition, issue, and state. I conclude that copies with the first setting of gatherings B-H and K are one edition (described here). Copies containing the second setting are a second edition, with three states corresponding to phase 2, 3, and 4 of gathering A (described here).

It seems odd that only some of the copies of the second edition say "second edition" on the title page, but that is the difference between bibliography and cataloguing. 

Do you agree?

Did you enjoy your taste of Bowers?


1Bowers, Fredson. Principles of Bibliographical Description. St. Paul's Bibliographies, 1987
2Randolph, Thomas, Aristippus (1630). 

Wednesday, April 8, 2020

The First Piracy. Part 4. Analysis

updated June 29, 2020
Part three of this series looked at gathering A of the first piracy of Hoyle's Whist. I described differences in the half-title, the title, and the advertisement. It would be useful for you to have that essay open in another window alongside this to be able to see all the photographs there. To recap:
  • The half title (A1r) is frequently not present in surviving copies. Ignoring one odd copy, almost certainly a later facsimile, two settings of type survive and initially there is no way to determine their priority. 
  • The title page (A2r) occurs in four variants, but sequencing is clear. Some type shifted between variants 1 and 2 leaving an oddly-spaced word "Treatise". Variant 3 improved the spacing in "Treatise". Variant 4 is like 3, but adds a line "second edition."
  • Some type shifted at the bottom right margin of the Advertisement on A3r. The earlier variant is is correct, while variant 2 has some inadvertent oddities. 
I have seen more than half of the surviving 44 copies and received reports on the others. There is regularity in how the four variants of the title page combine with variants of the half title and the advertisement:

Variation in gathering A

I said I have data on 44 copies and the chart shows only 43. One copy at the Bodleian Library, shelf mark Jessel e.640 lacks gathering A altogether, so one cannot say which title page it might have had.

With the title page variant 1, we always have the first variant of the half title and the first variant of the Advertisement. This includes 31 of the 44 surviving copies of which 6 no longer have a half title.

It seems that after these were printed,the half title was reset to variant 2. At the same time, there was some rightward movement of type in "Treatise" at the top of the title page, but no change in the Advertisement. I would have more confidence in the regularity if there were more than 4 copies of variant 2.

Why was the half title reset at this point in printing? One possibility is that the compositor needed the type from the half title to print some other work. The shop would have had fewer pieces of the larger letter forms that were used in the half title than those used for the text. You could imagine that the printing was interrupted, the type for the half title distributed and used elsewhere, and then the half title was reset. In the course of making this deliberate change, accidental changes were introduced on the title page while the type was unlocked. Speculation, to be sure, but it does account for the variation we see. 

The third variant of the title page moves the "T" in treatise to a more pleasing position, perhaps a deliberate change. At the same time, there is accidental type slippage in the lower right margin of the Advertisement. Only three copies survive with the variant 3 title page.

Finally, variant 4 adds the line "second edition" to the title page. There are five survivors from with two missing half titles. The motivation for this is, I think, clear. The printer/publisher James Mechell wanted to make his piracy appear to be a popular seller. Likely, he wanted to have a second edition to compete with the authorized second edition published by Francis Cogan.

To repeat, the counts vary greatly between variant 1 on the one hand (31 surviving copies) and the later variants (4, 3, and 5 survivors) on the other. That makes conclusions about the later copies more tentative than those about the earlier ones.
Note that by looking at many copies, we can determine the priority of the half titles--variant 1 always occurs with the earliest title page and variant 2 with one of the later title pages. This is a matter of how the book was printed--in gathering A the half title was in the same forme as the title page.

Similarly, we can play the same game with the variants in other other gatherings and determine priority. The rules are a bit different. Here are the counts of setting 1 and 2 for each phase of printing of gathering A:

Setting of later gatherings grouped by title page variant

In general, you can see that title page variant 1 has the setting 1 of all the gatherings, while title page variants 2-4 have setting 2. In part 2 of this series of essays, where I showed to settings of each gathering side-by-side, I conveniently put setting 1 on the left and 2 on the right.

We were able to determine the priority of gathering A (and in particular of the half title) because all the pages (the half title, title, and Advertisement) were in the forme at the same time. The priority of the title page and Advertisement were clear and we were able to infer the priority of the half title.

The story is a bit different with the other gatherings. They were not in the press at the same time as gathering A; indeed each was a distinct unit of printing. It was the binder who stitched all the gatherings together to form a book. It is logical that early versions of gathering A were sewn together with early versions of B, C, D, and so on, and that lets us infer the priority in printing the later gatherings. What I have called setting 1 was printed before setting 2.

The data is quite consistent, but there are exceptions. For example a variant 1 title page at the Newberry Library (shelf mark V 1639.42) has a second state gathering E, while another (reported to me by a bookseller) has second state gatherings B and K. There is also an odd copy from title page variant 4 at the Bodleian Library (Jessel e.641) which has the first setting of all gatherings except B.

Recall Jessel e.640 that lacked gathering A, so we couldn't identify which printing phase it belonged to. It has second states of all later gatherings.

Another observation is that for each later gathering, about 32 of the 44 copies have first settings. This suggests that Mechell decided to increase the print run by roughly a third when it was nearly done.

This was a quite detailed look at one book. With many surviving copies and many variants, we could infer a great detail about its printing history. If you enjoyed this, you may want to look at a much, much more sophisticated example: see Stephen Tabor, "James Shirley's Triumph of Peace: Analyzing Greg's Nightmare" in Studies in Bibliography 60:107-212 (2018) available online.

There is one issue I'd like to take up in the next essay. How does the printing history we have seen work with the cornerstone concepts of bibliography: edition, issue, and state?

Monday, April 6, 2020

The First Piracy. Part 3: Variants in Gathering A

updated June 29, 2020
From parts one and two, the story so far:
  • The first piracy of Hoyle's Whist is an octavo gathered in fours, and collates 8o: [A]4 χ2 B-M4.
  • The printing would have started with the B gathering and continued through to M. Gatherings A and χ would have been printed last. 
  • There are two settings of type for gatherings B-H and K. Both settings were printed at the shop of James Mechell. From the evidence presented so far there is no where to determine which might have been printed first.
  • There is one setting of type for gatherings A, χ, I, L, and M. 
Even though there is only one setting of type for gathering A, there are interesting variants to be discussed in this essay.

First a short digression that will make sense only to those who have studied physical bibliography and imposition schemes.
<biblionerd> An octavo in fours can be imposed in two way. In "work-and-turn", all pages of one gathering are set in a forme. It is printed (worked) on one side, the paper is flipped (turned), and the other side printed. You end up with two copies of the same gathering.

In "two sheets worked together" you set half of the pages in one gathering and half the pages of another in one forme. After printing one side, you distribute the type and set the other pages of each of the two gatherings. After printing the second side, you have one copy of each gathering.

The Webster piracy was printed work-and-turn1 so that a setting of type corresponded exactly to a gathering. Our story would be much more complicated if the book were printed the other way. In an early draft of this I showed the two imposition schemes and how I was able to determine that it was printed work-and-turn. I left it out of this final version because (a) it distracts from the story I'm trying to tell; and (b) only a small handful of you would care. </biblionerd>
On to gathering A.

Leaf A1r is the half-title used to identify the book before it is bound. Binders would often remove the half title before binding, in part because it is inessential and in part because binders would sell waste paper back to the paper mills.2

Of the 44 copies I know about, 12 lack a half title. The most common half title, appearing in 25 copies is this:

A1r variant 1

Notice the word game is in small caps--that is a large capital G followed by smaller, but uppercase letters: GAME. The second most common half title appears in 6 copies:

A1r variant 2

There are lots of obvious differences from the first variant, but the clearest is that "game" is in all capitals without the larger initial "G".

For those of you keeping score at home, I've accounted for 43 of 44 copies. Here's a final variant appearing in one copy only:

A1r variant 3
Vanderbilt University Special Collections
GV1277.H87 1743a c.1

The final variant looks odd. The type does not appear to be from the 18th century. The book is from the United States Playing Card Company/George Clulow collection at Vanderbilt. Clulow had many of his books rebound by fine 19c binders such as Ramage and Root & Son. This copy was bound by Root and like many Victorian bindings of the era, this copy is tightly bound. I cannot determine if the half title is conjugate to A4. My guess is that Clulow asked the binder to insert a facsimile half title to replace one that was missing. The practice of "perfecting" a book by adding missing pages was a fixture of 19c book collecting. As I said...this is a guess, but I will make no further reference to this oddball half title and treat as though it were missing.

Surprisingly, there are four variants of the title page, leaf A2r. While there is no immediate way to determine which of the two half titles was first set in type, we can sequence the four variants of the title page. Consider these two title pages, which as I will demonstrate, are the first and fourth variants.

A2r variants 1 and 4

These are the same setting of type. For example, look closely at the second paragraph beginning "Calculations, for those..." Note the large gap between "Bet" and "the" in both variants. On the line below, look at the gap in the final "the."

There are some differences. The "R E A" in Treatise on line 2 have drifted right in variant 4. Note the "A" appears directly over the "G" in "Game" in 1 but to the right in 4. More important are the added lines, a rule and "The SECOND EDITION" at the bottom of the right-hand example. The insertion has changed the vertical alignment throughout. Variant 4, the so-called "second edition", was printed after variant 1.

As a double-check that the type was the same, I photocopied variant 4 on a transparent sheet and laid in on top variant 1. With the changes in vertical spacing, I had to take many photos to get the type to align:


Let us compare variants 1 and 2:

A2r variants 1 and 2

The main difference I see is that "TREA" in "TREATISE" on line 2 have shifted to the right. There is a bit of type movement along the right margin in the final lines. The rest looks unchanged.

A2r variants 2 and 3
variant 3 from Copisarow collection, picture by Alexandra Ciucu

As we move from 3 to 4, the "T" in "TREATISE" has moved back left. The spacing looks better, though still imperfect. Perhaps the change was deliberate.

A2r variants 3 and 4
variant 3 from Copisarow collection, picture by Alexandra Ciucu

The "T" stays left, but the edition statement is added and everything is realigned vertically.

I conclude that there is one setting of type with changes in the press, some accidental and others deliberate. The movement in type and the addition of "Second Edition" lets us sequence the four variants as I have numbered them.

We can also discern sequencing in the Advertisement on A3r:

A3r variants 1 and 2

These two pages will give you a chance to read the beginning of the amusing Advertisement. They are the same setting of type, but something has happened along the right margin at the bottom of the page. In the word "Manner" at the end of the third line from the bottom, you can see the "r" has drifted up in the right-hand copy. One line down, the semicolon after "Undertaking" has disappeared and is replaced by a hyphen, which came from the "pre-" in the line below. Down another line you see that "pre-" picked up the final "l" from the catchword "vail" which has turned to "vai,". Where the comma came from, I cannot say.

The correct page, the one on the left, was printed first and the type shifted in the forme at some point, introducing the anomalies seen on the right. 

In the next essay, we analyze what we have seen.


1To learn more about the two imposition schemes and the challenges in distinguishing them, see Povey, "On the Diagnosis of Half-sheet Impositions," The Library, 5th ser. 17 (1962) 197-212, reprinted in Jones, ed., Readings in Descriptive Bibliography. Kent State University Press. 1974.

2Carter and Barker, ABC for Book Collectors. Eighth Edition. 2004.

Friday, April 3, 2020

The First Piracy. Part 2. Variants

A quick recap from the previous essay, "The First Piracy. Part 1. Overview and Structure".
  • The book was printed by the piratical James Mechell. 
  • It is an octavo gathered in fours, collating 8o: [A]4 χ2 B-M4.
  • 44 copies survive, by far the largest number of any of the early Hoyles.
  • The printing would have started with the B gathering and continued through to M. Gatherings A and χ would have been printed last. 
I also mentioned that there were variants among the copies. Let us look at the first page of text from two different copies. The text begins on page one, leaf B1r. B1 is the first leaf in the B gathering and "r" refers to recto, the right-hand page of an open book. Page two would be B1v with "v" for verso.


The text is identical, word-for-word, but there are many typographical differences:
  • The example at left lacks a page number; the right is numbered "[1]".
  • The drop-title "A Short Treatise on the Game of Whist" is completely different. Look at the line breaks, the use of capitals and the use of italics.
  • On the last line, the word "One" is capitalized in the first example but not in the second. 
I would be easy (and tedious) to find dozens of differences in spacing or particular pieces of type:
  • On the left, the "A" in "Author" on the first line of text appears above and slightly to the left of the "d" in "did" on the line below. On the right, it appears above the space betwee "tise" and "did".
  • Look at the "e" in "Attention" 4 lines from the bottom. On the left, it seems to me that there is a break in the top part of the letter suggesting some damage to the type. On the right, the "e" looks fine. For a different broken "e", look at "Payment" on the last line. 
Clearly these pages are different settings of type.

If you picked up a random copy of Hoyle's Whist, printed for W. Webster (1743), you could find either of the two variants. That is odd, even in the hand press era. Many questions come to mind. Why did this happen? Which was printed first? Were both printed in the same shop? I'll return to the "why" question at the end of this essay. The question of priority will have to wait for a later essay--it requires a lot more evidence and discussion.

We can conclude with reasonable certainty that both settings were printed in the same shop. Note the identical woodblock ornament at the top of both settings. Ornaments are lovely decorative elements from books of this period, and can be useful in identify anonymous printers. The fact that both pages share an ornament is a strong indication that both were printed in the same shop. It is theoretically possible that the ornament was loaned from one printer to another and each printed a page, but that strains credulity. Both were printed in the shop of James Mechell.

We could have as easy compared copies of page 2:


Differences include:
  • A different break between lines 2 and 3: "upon Pay-" versus "upon".
  • Different spacing between paragraphs.
  • A decorative woodblock ornament on the left, versus a line of type ornaments on the right. 
  • Below the decoration, the type is different in every way--the use of capitals and italics, line breaks, the two-line initial "I" on the right, different catchwords...
Let us see two settings of a page from gathering C:


Some of the many differences are:
  • The spacing before and after the page number.
  • The woodblock ornament at the top
  • The line break in the chapter title and the spelling of "observ'd" versus "observed".
  • The position of the signature mark C2 with respect to the text.
And D:


Some differences:
  • The spacing before and after the page number.
  • The woodblock ornament at the top.
  • Line breaks in the chapter title.
  • The capital S in "Suppose is a different piece of type in the two examples. 
I'll skip gathering E (which also has two settings), but move on to samples from F, G, and H:


You can make your own (long) list of difference by now. I let my eye run down the right margin and note differences in the line breaks beginning on the fifth line. The woodblock tail pieces differs as well. The next two examples are an exercise for the reader:



There is only one setting of type for gathering I; gathering K has two, and gatherings L and M have one only. As mentioned above, when there are two settings, we are not yet prepared to discuss which of them might have been printed first.

Gathering A with the half-title, title page, and Advertisement has only one setting of type, but there are fascinating and revelatory variations of a different sort. Please be patient until the next essay.

Why are there two settings of type of some gatherings and only one of others? Here is how I envision the workflow in Mechell's shop. The compositors--as discussed below, I suspect there were at least two--started setting the type with the text in gathering B. As each gathering was finished, the pressmen printed a fixed number of each sheet. The compositor(s) would then distribute the type and beginning setting another gathering.

When the book was about two thirds finished, Mechell decided to increase the print run, expecting more demand for the book than he originally supposed. The pressmen printed a larger number of I, L, M, A, and χ. Then, the compositors went back to reset gatherings B-H and gathering K. Enough copies of the resettings were printed off to make complete books for the larger print run of the later gatherings.

Why do I think there were two or more compositors? If there were only one, there wouldn't be the oddity that the earlier gathering I had only one setting while the later gathering K had two. I suspect that one compositor was setting gathering I and another gathering K. Gathering K was printed and the top distributed while I was still in the press. At the point Mechell decided to increase the print run; therefore I did not need to be reset, while K did. Speculation, yes, but it explains the pattern of reset gatherings.

It is amazing and interesting how much we can learn Mechell's thought process and workflow by looking at lots of copies of Whist. 

Next: Gathering A

Wednesday, April 1, 2020

The First Piracy. Part 1. Overview and Structure

(Portions of this essay appeared in a previous post.)
I've said a lot about the first piracy of Hoyle's A Short Treatise on the Game of Whist in this blog, in a published article, and in a talk I have given two or three times. There remains much to say about the physical book. There are variants among the surviving copies that reveal much about its printing history and challenge the bibliographical concepts of edition, issue, and state. It will take me several essays to discuss the variants and their implications. A theme throughout is the bibliographer's mantra: examine as many copies of a book as possible.

First some background. 

Hoyle published A Short Treatise on the Game of Whist in November 1742. He sold it privately to his whist students for the high price of one guinea. After selling out the first edition, Hoyle sold the copyright to bookseller Francis Cogan for 100 guineas on February 3, 1743. Cogan must have expected to sell the book for the same guinea that Hoyle charged and took a slightly marked-up copy of the book to printer James Mechell. Before Mechell printed the second edition for Cogan, he printed copies, lots of copies, to sell for his own profit.1

The "Webster" piracy

The piracy, pictured at left, omits Hoyle's name from the title page, attributing the book to "a Gentleman." The imprint "Bath printed, and London reprinted for W. Webster near St. Paul's" is fictitious. The book was never printed in Bath and Webster is a name invented to disguise Mechell's identity. The piracy was advertised in the General Evening Post of February 19 at a price of two shillings, less than a tenth of the one guinea that Cogan intended to charge.

It was not until the first week of March that Cogan published a second edition, matching the pirate's price of two shillings. In April, Cogan obtained an injunction against Mechell, James Watson, a second printer who pirated Whist, and seven booksellers who sold copies of the piracies.

I was not kidding when I sad that Mechell printed lots of copies. No records survive indicating the size of the print run, but I know of 44 surviving copies, by far the most of any early Hoyle. With only four known copies of the first edition, the piracy is the earliest Hoyle obtainable. Other Cogan editions of Whist survive in small numbers: eight copies of the second edition, seven of the third, ten of the fourth and seven of the fifth.

Let us start by looking at the physical book. The most important element of a book description is the collation statement which describes its structure. The piracy collates 8°: [A]4 χ2 B–M4. The formula means that the book was printed as an octavo (eight leaves or sixteen pages to the printed sheet) and assembled in gatherings of four leaves or eight pages. Gathering A is unsigned (as indicated by the brackets); B through M (omitting J, as is typical for books of the period) are signed. The symbol "χ" is used for an unsigned gathering in the middle of the book and the superscript "2" indicates that there are two conjugate leaves, that is a single piece of paper folded to make two leaves or four pages.

A stab-sewn copy, never bound

When I wrote the "Pirates" article, all the copies I had seen were tightly bound and I couldn't tell whether the two leaves between gatherings A and B were conjugate. I gave the more conservative collation formula 8°: [A]4 (A4+2) B–M4 indicating that the two inserted leaves were singletons, that is separate pieces of paper. In 2012, I got the copy pictured above. It is in completely original condition, unbound with the original stab sewing. Even though the pages are a bit curled, it is a delightful survival that reveals the book's structure.

The photograph of the bottom of the spine below, makes it clear that the two leaves between the A and B gatherings are a single folded sheet rather than two single leaves.

χ2 not A4+2

One would expect the printer to begin setting the type with gathering B where the text begins, and proceed through the end of the book. Gathering A, and here χ would be printed last. The first leaf A1r is the half-title pictured above, and A2r is the title page, also picutred. χ1 and χ2 contain the table of contents, obviously printed last.
Leaves A3 and A4 contain a curious "Advertisement" in the form of a "Letter from a Gentleman at Bath" which purports to describe the publishing history. The author describes losing "a considerable sum of money one night at [whist]." He concluded that he was beat by superior skill and found that there was "a treatise on the game of whist lately dispersed among a few hands at a guinea price." He obtained a copy "with no small difficulty" and learned he "had heretofore been but a bungler at this game." He "applied to a stationer who offered to make [him] a present of half a hundred of them, provided [he] would allow him to print a few more for his own use."

Well, that's not quite what happened!

The next essay will begin to look at variants.


1A fuller account is in my article "Pirates, Autographs, and a Bankruptcy: A Short Treatise on the Game of Whist by Edmond Hoyle, Gentleman" in Script & Print, 34 no. 3 (2010): 133-61, available for download.

Friday, December 20, 2019

2019: The Year in Collecting

updated June 4, 2020

2019 was a good year for me as a collector. Since my Hoyle collecting is far along, when I find something I don't have, it must be rare. I'll look first at the best of the new non-Hoyles, then the Hoyles, but first a story.

In the summer of 2017, I took David Pearson's class in Provenance research at the Rare Book School in Charlottesville. We learned about forms of evidence to understand book ownership: inscriptions, bookplates, bindings, heraldry, catalogues, and more. At a break, I told David that I had never marked my books in any way. As I described my collection, he said, "You have some really good books, but most of your books are interesting only because they're part of your collection. If you don't identify your books, that interest will be forever lost." His comment resonated--for example, I have a lot of whist books that are not of great value individually, but I do have 300 of them that together tell the story of the game. He convinced me that I should identify my books.

I contacted some friends who are letterpress printers and came up with the bookplate pictured below. The woodblock ornament is taken from the first edition of Hoyle's Backgammon. Add a border and my monogram (which David described as a mangled Roman numeral), and voila! My friends had a polymer plate made with many instances of the book plate in three different sizes.They printed and cut out lots of plates.
Levy bookplate
Polymer plate on the press

It has been a huge project attaching the plates. In addition to pasting them in the books, I've used the opportunity to make sure that my electronic and paper records are in sync with the physical books. Disclosure: they weren't! I've also added location detail so I can print a shelf list. Finally, I've also had to decide which books are part of the collection and which are not (such as modern books on contract bridge or backgammon).

I have plates in roughly 800 of the 1200 book collection so far. Work continues! I am using a rice paste that is soluble in water, so the process is reversible. If my books are important as a part of collection, they are now mostly identified.

On to the new acquisitions.

I have a weakness for books in manuscript. We tend to think that once moveable type was introduced, all books were printed. In fact manuscripts overlapped with printed books for a long, long time. For an extraordinary example with a great story, see my essay "The Left Hand of Bougy..." The manuscript on Quadrille from 1725 is, like the Bougy manuscript, copied from an edition of the Académie des Jeux, the French gaming anthology that appeared in one form or another for generations.

1732 Quadrille

Another weakness. And another book on Quadrille. I've written many times how I love books in original unsophisticated bindings. See, for example here, here, here (second from the top), and here. The book on Quadrille at left predates Hoyle; I mention it in the essay "Piquet and Quadrille Literature." I acquired the copy at left at auction this year. It is a pamphlet that has never been bound and retains the original stab sewing. Lovely!

And now the Hoyles.

Whist, Lisbon (1753)

This year brought a copy of the first Portuguese translation of Hoyle's Whist, Lisbon, 1753. When I wrote the essay "The First Translation of Hoyle," everyone thought that this was the earliest translation. Since then, a 1751 translation turned up, as I wrote in the essay "A French Discovery." So Do Jogo do Whist is the second translation of Hoyle. Mine is one of two known surviving copies.

What is remarkable about the book is that it is the first I know to illustrate the use of tokens for keeping score at whist. In the essay "Hoyle's Scoring Method and Whist Counters," I discuss whist scoring tokens. I suggest in that essay that the first mention of scoring with tokens was in the 1791 chapbook Short Rules for Short Memories at Whist by "Bob Short" (Robert Withy).  I still believe that to be the case in English, but I've since found scoring discussed in three 18c Portuguese editions (1753, 1768, 1784), and a Russian edition published in St. Petersburg (1769). The Russian edition purports to be a translation from a French edition, but the tokens are not illustrated in any French version I have seen.

Whist scoring tokens (1752)

The two new Hoyles in English are both reissues of books I have already, but with cancel titles. The first is discussed in "Reissues of Mr. Hoyle's Treatises (1748-1755). With this 1755 reissue, I have four of the five different issues.
Mr. Hoyle's Treatises (1755)

The final example is a Dublin reissue of the Polite Gamester from 1783. It was originally published by Thomas Ewing in 1772. Ewing died in 1775 or 1776 and James Hoey took over his stock. Hoey reissued the book with a cancel title in 1776. See the essay "Every Cancel Tells a Story, Don't It (part 1)" for pictures and more detail.

I can't say why Hoey put a new title page on the book in 1783. The imprint is the same as is his address, 19 Parliament Street. Probably, he wanted to make the book look more current with a new date.

The Polite Gamester (1783)

It's a bit beaten up, but the only other recorded copy is in the John White collection at Cleveland Public Library (discussed here), so I have no complaints. 

Best wishes for 2020!