Thanks batgirl, for lots of illumination as ever
algebraic notation

I have Dutch chess books from the 50s and they're in algebraic. I don't know how it was pre-war, but I suspect it was algebraic too. I've always viewed descriptive as something you see in old American chess books, nothing more.
Bit like imperial vs metric. You use rods per hogshead, we the normal stuff
Edward Winter's Chess Notes should contain more historic stuff. A quick search at least shows up http://www.chesshistory.com/winter/extra/1924.html about chess in 1924, quote:
The same month F.P. Betts of London, Ontario had a letter published on page 204 of the American Chess Bulletin in which he denounced a book by Edward Lasker (unnamed, but presumably Chess and Checkers The Way to Mastership) for using the ‘extraordinary’ algebraic notation. ‘This mania for improving the established chess notation by freak innovations seems to entirely overlook the fact that there is an immense volume of chess literature centuries old in which the old and accepted notation is enshrined, the value of which would be greatly impaired if this novel jargon should become the vogue.’
So this debate isn't new, not even in the US.

Thanks for the head's up on R. F. Green's Introduction to Chess in the Chess Player's Handbook !! Here's what he had to say:
p. 500
R.F. Green is not the author, but an editor of a reprint of Howard Staunton, The Chess-Player's Handbook (1847). See the full text of an 1890 edition at Google Books. I have on my shelf a paperback 1995 reprint edited by Raymond Keene. At least in this case, he does not make an effort to pass off Staunton's work as his own. Indeed, thankfully, a one page Preface offering a tidbit of historical information is the sole cause of Keene's name on the cover as editor. The pagination of Green's 1890 version and Keene's 1995 is the same. Both are identical to Staunton's 1864 edition also digitized by Google Books.
It is worth noting that long algebraic is fully formed in Staunton's book, while the descriptive notation many of us learned in our youth had not yet developed it's efficient form. Staunton writes P-K4 as "P. to K's 4th."
OTOH, I think the texts by Greco, Ruy Lopez, Damiano, ... have a version of what would come to be known as descriptive. Perhaps we could say with some accuracy that a long and cumbersome form of descriptive is older than algebraic, but that algebraic required fewer changes to reach its modern form. The "old" descriptive notation is principally a twentieth century phenomenon in English speaking countries. But, a longer version of description is the origins of chess notation.

Ziryab, in fairness to R.F. Green there is another book solely by him called Introduction to chess and that may be what was referred to earlier by Jenkins 12, I have an updated copy of it by J Du Mont dated 1938. I'm not sure Green ever tried to take credit for Staunton's work :-)

Green doesn't take credit for Staunton's work in the 1890 edition. It's quite clear that he brought Staunton's text back into production, and he added an appendix (pp. 519-544) that includes an alphabetical list of openings and a bibliography of chess literature. Green was editor of British Chess Magazine, and as such made many contributions in bringing the work of others forth.
Keene, on the other hand, has a reputation as a plagiarist.
I should clarify my earlier statement that Green's and Keene's editions of Staunton's works are the same. The pagination is the same up to the end of Keene's text. But, Keene does not include Green's appendix. Nor should he. Both the opening reference and bibliography would have been terribly out of place in 1995.

Ziryab, I misread your post, I thought the *he* referred to Green not Keane :-) Someone bought me a copy of that republished handbook with Keane's preface as well. It's hard reading...the Staunton book I mean.

I agree that some labor is needed to read Staunton, but I think it's worth it, at least for historical interest. For chess improvement, more recent books are better focused. Most nineteenth century texts seem prosaic by today's standards, but that's part of their charm.

I was researching this a bit and came across an 1821 edition of Philidor's l'analyse that used algebraic notation:
as opposed to the 1749 original that used descriptive:
Cette édition (1821) utilise la notation algébrique : dans la première, Philidor avait manifestement refusé de prendre en compte l'évolution introduite par son rival Stamma.

That's illuminating Batgirl. Thanks.
Any info on publisher? Who is responsible for the alteration? Can you translate the French for those of us ignorant of civilized language?

as requested, Ziryab (I have a bat avatar although I'm not batgirl, but then you didn't address that question specifically to her).
25. The Pawn of Insane of the king a step The Pawn of insane of the lady a step
This edition (1821) uses the algebraic notation: in the first, Philidor had obviously refused to take into account the evolution introduced by its Stamma rival.
(translated using babelfish)

The French statement says, loosely: This edition uses algebraic notation: in the original, Philidor had manifestly refused to utilize those changes that had been introduced by his rival Stamma.
Chronologically, Stamma published his book in 1745. Philidor beat Stamma in their London match in 1747 and Philidor published his own book in 1749.
Here is the title page of the 1821 edition:
The French on the title page mentions the book contains the new agebraic notation, 42 positions as well as explanations of the movement of the pieces and of chess terminology.

To summarize, so far we've learned that algebraic notation probably started with the Syrian player/teacher Philipp Stamma. While Philidor and others endorsed a descriptive notation which was almost totally adopted by England and America, other countries, particularly Germany and Russia (and seemingly the Netherlands) who at that time were approaching chess in the most scientific fashion, adopted algebraic notation. Descriptive notation remained the mainstay of America (and England too, it seems) until the late 1970s - early 1980s when algebraic notation, whether by the insistance by FIDE or by the advent of computers, or both, started gained popularity and eventually superceded descriptive notation completely.
Is that a fair assessment?

The French on the title page mentions the book contains the new agebraic notation, 42 positions as well as explanations of the movement of the pieces and of chess terminology.
Well, with nothing else to go on, this source alone would suggest that algebraic had become popular in France in the early nineteenth century. That offers a partial answer to the original poster's question, and answer not rooted in Anglocentric assumptions.
Of course, this one text is not all that we have. But, it does seem an importnat story--changing Philidor's notation over his dead body so to speak.
RE: Babelfish:
"25. The Pawn of Insane of the king a step The Pawn of insane of the lady a step."
Fou should perhaps be translated fool, rather than insane: "The pawn of the king's fool one step; the pawn of the queen's fool one step." I've heard it said that the French call the bishop a fool, although insanity might be just around the corner.

To summarize, so far we've learned that algebraic notation probably started with the Syrian player/teacher Philipp Stamma. While Philidor and others endorsed a descriptive notation which was almost totally adopted by England and America, other countries, particularly Germany and Russia (and seemingly the Netherlands) who at that time were approaching chess in the most scientific fashion, adopted algebraic notation. Descriptive notation remained the mainstay of America (and England too, it seems) until the late 1970s - early 1980s when algebraic notation, whether by the insistance by FIDE or by the advent of computers, or both, started gained popularity and eventually superceded descriptive notation completely.
Is that a fair assessment?
Yes

At the risk of muddying the stream, I offer this image of the notation in an 1806 text:
And the cover of the text:

I have to agree that Pawn to King Four sounds much more assertive and impressive than e4. But given the other trade offs I'd much rather hear e4. I can remember when there was a big push to adopt Algebraic and I was behind it all the way. When there are only six or seven chess men left on the board and it's Kt-QB7 you have to think about where the Queen's Bishop was at the start of the game and which side of the board are you talking about? Giving each square a unique name is a vast improvement. Kt-QB7 becomes Nc7 if it's White to move and ...Nc7 if it's Black to move. Algebraic was such an enormous improvement that I really don't understand why it wasn't adopted much earlier. The only thing I don't like about algebraic is that it has a sleight favoritism towards White, ergo; if you're talking about White's Third rank - that's a3 thru h3. If you're talking about Black's Third rank it's a6 thru h6, so it's actually the sixth rank. But that's minor. I rejoiced when USCF officially adopted Algebraic.

One of the reasons I posted this was that I found it ironic that "Cette édition (1821) utilise la notation algébrique : dans la première, Philidor avait manifestement refusé de prendre en compte l'évolution introduite par son rival Stamma" was written about a treatise in which Philidor's notation was altered to be more like Stamma's notation.
I'm not sure this indicates that the French were devoted to algebraic.
"Philadelphie: Chez J. Johnston, Libraire-Editeur"
I would imagine that this means it was published in Philadelphia?
Except . . . Willard Fiske listed all the chess books published in America prior to 1860. He lists the first edition of Philidor in America is an 1826 English language reprint of an 1819 English edition (also reprinted in England in 1824) in which the "moves are written out in full length" (i.e. descriptive).
go figure
As I recall, it was at the insistance of the FIDE, in the late seventies and early eighties, that it be used, in their tourneys sanctioned in Europe, that the algebraic (german) system became frequent. And we all know how wonderful the FIDE has been for the game of chess don't we?
That's, also well before personal computers were more than an oddity folks. Most computer afficianados back then were amazed that they could play tic-tac-toe on one at the time.
At the time, and it probably wasn't the cause, it appeared, from here, to have been a large, group, political inferiority complex, striking a final blow for the end of anything to do with that terrible system across the ocean. You know from that place where that terrible Bobby, who wouldn't go along with a rule fix favoring the Russians just to get along, came from. That bad man who forced the indignant sponsors to pay prizes that allowed the good chess players to almost make a living.
Unfortunately for those of us who learned the moves 50+ years ago, algebraic is difficult. And, this goes for many of the players I've known for years. We just don't think in algebraic. I still think in descriptive and then translate. And, even though descriptive is still an approved form of notation for all OTB tourneys, bland algebraic is more frequently used.
As far as ease of use, it's a coin toss. I automatically know which file, and side of the board, is K or Q and which kNight, bishop & rook go with each. Also, they're the same from both sides of the board. If someone finds Nf3-d4 easier to write than QN - K4 and hit the clock, I fail to see how.
As far as use, just think back to the egos involved at the end of the 19th and the early part of the 20th century. I don't think anyone in that group [whether Steinitz, Lasker, Alekhine, Capablanca etc.] was going to let his side of the board be less than the "first rank" or concede from an eighth rank, psychologically at least, and inch or a centimeter as the case may have been.
One of the misfortunes [Or perhaps good fortune for them] is when I lend an older book to someone who has only algebraic learning. They have difficulty reading it and have to learn descriptive. It's rather like the difficulty for an English speaker learning another language or someone who has only one of the Romance languages learning English with all it's double and triple meanings for the same word spellings depending on how it's used.
But to read a book by Emanuel Lasker, Fred Reinfeld, Al Horowitz or even that "evil" Bobby, as it was written, it's worth learning.