Data Format

Notes

First Editions

Officially a first edition is the first printing of a book, done from the original setting of type. The collectibility of the first printing of the first edition was established in the early days of printing, when the lead type used in the presses would quickly wear away, compromising the readability of the book being printed.

Technically, all of the copies of a book printed from the original setting of type, at one time or over a period of time, with no major changes, additions or revisions, would also be considered a first edition. Minor changes, such as the correction of some misspelled words, or the addition of a dedication, or similar very minor alterations, may be made and the revised copies are still considered as part of the same edition, simply being described as different states or issues. It is in this sense that a "first edition" is of importance to researchers as it often represents the "original form" of a text before later alterations were made.

In the collecting world, however, the term "first edition" has taken on a life of its own, with which can be associated huge sums of money (compare the cost of a set of Harry Potter "firsts" with the cost of buying exactly the same text in mass-market paperbacks which, if set from the same computer files, would also, technically, be first editions). Because of the money involved, much heat can be generated over what is or is not the first edition of a given book.

The biggest complicating factor is that there is no easy (or agreed) way of defining when a particular edition was published. While it is usually fairly easy to identify the "official publication date" for a book (i.e. the date assigned by the publisher), it is often argued that this date is inaccurate given that books are often available in bookshops some weeks prior to this date (particularly in the US and particularly around Christmas). Unfortunately, there is no reliable way of establishing when a book is first available in bookshops (nor any agreement as to whether the sale of a single copy in a specialist bookstore is sufficient, or whether it needs to be available in general booskhops throughout the country concerned.

This, in turn, is rapidly becoming academic given the decline of bricks-and-mortar bookstores and the growth of online retailing. There have been some moves to use the date specified by Amazon as the deciding factor, but Amazon hardly have a good track for accuracy in such matters. Others have proposed using the dates of mentions in Publishers Weekly but, at best, this only covers a small subset of books published.

Despite the inherent inanity of the whole situation, these is still a lot of money involved, leading to some truly bizarre behaviour at times:

At the height of the Stephen King boom, an attempt was made to make his new book available on the same day worldwide. As a result, one "enterprising" dealer bought up a large batch of the New Zealand edition which he then sold as the "true first edition" because the bookshops opened there earlier than anywhere else in the world (as a result of time zones). One result of that was that some later "high profile" books were made available at exactly the same time worldwide, with bookstores opening at 3AM to satisfy the local addicts.
In the late 1980s, Easton Press published a series of limited "first editions" of new science fiction books, which often appeared months after the trade edition. When Locus started listing these as reprints, they were threatened with a lawsuit by Easton Press and forced to list them as first editions, even when everyone knew they weren't.

Some further complications include:

Most wealthy book collectors are American and are disinclined to accept any non-US edition as being a "real" first edition, such that "first US edition" will often attract higher prices that a UK (or Australian or Canadian) edition that appeared months earlier.
Similarly, "book club" editions are generally regarded as of considerably less interest, even if available simultaneously with the trade edition.
Similarly, a hardcover edition, particularly one with a strictly limited print-run, is generally regarded as of more interest than any other edition.

As a result, in the late 1980s, Charles Brown (then editor of Locus) laid down a number of guidelines for the contents of this field:

If the US and UK publication dates are the same month, then each is a first edition, but the US publisher gets listed first in reprints.
If the UK publication date precedes the US publication date by one month, then each is a first edition, but the UK publisher gets listed first in reprints.
An SFBC edition is only a first edition if it precedes the trade edition by at least one month.
The Easton Press series is always called a first edition, but never appears as such in the reprint data.

As a result of the above chaos it is probably fair to say that, outside of the Locus columns, little attention is paid to whether or not a book is deemed to be a "first edition".

First Appearance for Items

In a well-ordered universe, there is little doubt about the first appearance of an item - it appeared in a magazine and then was reprinted in a collection. Real life, however, is much messier and the same item may appear "simultaneously" in multiple magazines and/or appear in a collection before (or simultaneously with) appearing in a magazine. As with first editions there are no clear rules for establishing precedence, but fortunately there is also much less interest in the topic so we can establish our own guidelines without worrying too much about people getting upset.

Simple Cases

The first simple (and obvious) rule is that the item dated the earliest is the first appearance - an item with a publication date of, say, April 1991 always takes priority over an item with a publication date of May 1991, no matter what the other circumstances. Even here, though, there is a degree of additional complexity:

If we know the month of publication for one and only the year for the other, then the former takes precedence over the latter unless the years match and the former states it is reprinted from the latter.
If one publication is monthly and the other more frequent, and for the same month, then the monthly takes precedence (e.g. an issue dated April 1991 is deemed to precede an issue dated April 5th 1991).
Magazines that have complex dates (e.g. Summer 1945 or 2nd June Issue, 1933) then the dates should be translated in the same way as on the book record. If the resultant date is identical to that of the "other" source then the other source, being more exact, is deemed to have precedence.
If both dates are identical, and one source is a "reprint edition" of the other then the "parent" magazine takes precedence.
If both dates are identical, and one source is a magazine and the other a collection or anthology then the magazine takes precedence.
Otherwise, a UK source takes precedence over a US source and both take precedence over any other country. If both sources are from the same country, or neither is from the UK or US, try to establish some relevant factor (e.g. if one is from the author's home country then let that take precedence) and if that fails toss a coin!

Incomplete Data

A complication arises when we have partial data about the original appearance, but more concrete information on a later printing. For example we may know that a story appeared in a collection which said that the story had previously appeared in a given magazine, but gives no details, or that it was first published in a certain year but doesn't say where, or possibly just know that other stories in that collection came from a particular magazine and guess that the same might apply to the current location.

In due course, some of these will be addressed by use of the Item Printing Records, but for now the best approach is to list the first known appearance and add an Item Note Record to describe what is known about the prior appearance, as in:

E0041A1~Crispin, Edmund~Lacrimae Rerum~ss1949EQMJun~~~Fen| Gervase~
E0041D1~probably first in {The (London) Evening Standard}.~

E0041A1~Crispin, Edmund~Lacrimae Rerum~ss1949EQMJun~~~Fen| Gervase~
E0041D1~first published in {The (London) Evening Standard}, date unknown.~

E0041A1~Crispin, Edmund~Lacrimae Rerum~ss1949EQMJun~~~Fen| Gervase~
E0041D1~supposedly first published in 1946, publication details unknown.~

Serials

In most cases, serials can be handled in the same way as ordinary items - if the serialisation starts in one magazine before it starts in another and finishes in the first before it finishes in the other, then the first clearly takes precedence. Similarly, if the serials coincide precisely, the same precedence rules apply as above. However there are rare cases where the serials overlap in such a way that it starts earlier in one magazine, but finishes earlier in the other. In most cases such serialisations are still identical (i.e. the same number of parts with the same contents) so each should be handled as individual items following the rules above. Thus the first few parts of a serial might be original in one magazine and the later parts original in a different magazine.

If the serial is then further reprinted in another magazine (or collection), then the magazine in which it finished first should be regarded as the original appearance, with an Item Note Record recording the other appearance, as in:

E0000A1~Smith, Fred~Exciting Story~na1920AMGMar+3~
E0000D1~also published in {Pearson's} from January to August 1920.~

In the extremely rare instance where the serialisations are different (i.e. different numbers of parts) then both should be listed as original in the magazines themselves but, again, any future reprints should indicate the magazine in which the serialisation finished first as the original appearance.

In all such cases, an Item Printing Record should be used to indicate the simultaneous printings.

Reprints of Serials Published as Books

In many cases, an item (typically a novel) that has been serialised in a magazine is (later) published as a book. If the book publication occurs before the serialisation starts, then clearly it takes priority and the serial is a reprint. If the book publication occurs during the serialisation then the serial parts should still be listed as if original to the magazine but an Item Note Record should be added to indicate the priority of the book, as in:

E0040A1~Pohl, Frederik~Jem [Part 5 of 5]~n.1980GALJul~
E0040D1~note that the completed book was published by St. Martin's Press in April 1979.~

A more common situation arises when a novel (or similar), which has appeared as both a magazine serial and a book, is subsequently reprinted in a magazine or collection or similar. Should the original appearance be listed as the magazine serialisation or the book publication?

The first point to bear in mind is that there are probably textual differences between the two appearances so there is, technically, a "right" and a "wrong" answer and if that is known then it takes precedence over the guidelines below. In most cases, however, it is unlikely that there will be any way of telling which version is being reprinted, in which case the following guidelines should be followed:

If the magazine (or collection or whatever) in which the reprint occurs is reprinting similar items from the original magazine or is reproducing a facsimile of the original magazine, then the magazine serialisation is listed as the original appearance
In all other cases after the book has been published, the first edition of the book is cited as the original appearance

Where appropriate, particularly in the latter case, an Item Note Record should be used to indicate the "other" publication, as in:

E0000A1~Burroughs, Edgar Rice~At the Earth's Core~n.1914ASWApr 4+4~~~Pellucidar~
E0000D1~book edition published by McClurg in 1922.

E0000A1~Burroughs, Edgar Rice~At the Earth's Core~n.1922MCG~~~Pellucidar~
E0000D1~originally serialised in {All-Story Weekly} from April 4th 1914 to April 25th 1914.~

Reprints of Multiple Items under a Single Title

In certain story papers, in particular, it was a fairly common practice for a group of stories, initially published individually, to be reprinted under a single title (rather like a fix-up novel), as in:

E A1~Conquest, Owen ,(hp:Hamilton, Charles #2)~Algy Silver's Pal~nv1918+ByFrdApr13+1~~~Rookwood~
E D1~originally published in two parts ("Algy's Pal!" & "Betrayed by His Chums!").~

The above format is not ideal as it implies the story was a two-part serial in ByFrd called "Algy Silver's Pal". Instead the constituent stories should be listed indvidually as embedded items, as in:

E    A0~Conquest, Owen ,(hp:Hamilton, Charles #2)~Algy Silver's Pal~nv~~~Rookwood~
E   _A0~Conquest, Owen ,(hp:Hamilton, Charles #2)~Algy's Pal!~ss1918+ByFrdApr13~~~Rookwood~
E   _A0~Conquest, Owen ,(hp:Hamilton, Charles #2)~Betrayed by His Chums!~ss1918+ByFrdApr20~~~Rookwood~

Generic Titles

A particular problem exists where the title of an item is a "generic" title that has probably been used on several different items - the most common of these being poems with titles such as "Song" or "Sonnet". Thus, for example, there are 7 poems by Charlotte Becker simply titled "Song" - 3 in Ainslee's Magazine, 1 in Everybody's Magazine, 2 in The Smart Set, and 1 in Snappy Stories. It is highly unlikely that all 7 are the same poem but, at the same time, quite likely that some are reprints. As there is no way of identifying which items are reprints of which, the convention is to list all 7 items with no publication details at all. This is not ideal, but is better than making assumptions about which are or are not original.

Note that, if the text of the relevant issue(s) is available then the first line of the poem should be appended to the generic title to differentiate it, as in:

Loveman, Robert~Song ("Back to the siren South")~pm1902AINNov~
Loveman, Robert~Song ("Flora is a famous flirt")~pm1904AINSep~
Loveman, Robert~Song ("I weep so often now")~pm1900+HpNMMNov~
Loveman, Robert~Song ("Over the sea we go")~pm1900+HpNMMAug~
Loveman, Robert~Song ("Sing it away")~pm1906AINFeb~
Loveman, Robert~Song ("The dark is dying, dying")~pm1904HPMOct~
Loveman, Robert~Song ("The dawn is a wild, fair, woman")~pm1901HPMAug~
Loveman, Robert~Song ("The dream is o'er")~pm1908AINJun~

Note that a similar problem occurs with anonymous items (particularly poems) that have the same title. Unless the title is really unusual it is unreasonable to assume that any two such are necessarily the same item so, if there is any doubt, the publication details should also be omitted from all such occurrences.