Regarding practice in the USA before Vatican II: that era was one where Irish and Irish-Americans dominated diocesan and parochial leadership in much of the most Catholic areas of the USA, and overall (with individual exceptions), the sung Mass tended not to rank highly on their priority list, shall we say.

The GR was taught and used in the Archdiocese of Milwaukee due to the influence of John Singenberger, who had been trained at Regensburg. Regrets cannot tell you its penetration as a % of parishes but the German parishes used it.

Your question is unclear to me. Are you attempting to contrast the Graduale Romanum with the Liber Usualis? What relevance does the USA have to Durufle?

I'm virtually certain that MD used the LU for the Requiem.

That his Requiem is based on the Gregorian chants as interpreted by Solesmes is undoubted. This chapter from J.E. Frazier's book on Durufle has plenty of detail on his opinion of chant and the Solesmes interpretation. If you're trying to find evidence that he used the Liber Usualis over the Graduale Romanum, best of luck to you, but I don't see what that proves.

Thank you, GerardH.
I'm the author of that biography of Duruflé. In it I did not pursue the LU-GR issue. I made no substantive reference to those books at all. It is only now, a quarter century later, that the matter has grabbed my attention.
When MD was a chorister, there was a palpable shift in the way chant was sung at the Rouen cathedral. In his memoir (1976) he called it a "true revolution." Compared to the previous method, he found the new practice truly compelling. He did not say so directly, but it is clear from the context that the Solesmes method was the key to that revolution. I'm confident that experience in his early teens led ultimately to his quest and success with the Requiem.
I'm also increasingly of the opinion that the cathedral made a shift at that time from one chant book to another. For reasons that I cannot explain, the matter interests me greatly. But I have found no interest in it among chant historians nor have I found anything relevant to it west of the Atlantic. I'm confident there's an answer in Rouen's archives.
The (North American) internet shows that writers overwhelmingly (but without documentation) believe MD drew his material for the Requiem from the LU. That opinion is as vast as it is pointless, in my view. I'm simply trying to track down the source of that opinion. (The LU was the only book known and used in the US before the council. The GR was completely unknown. I don't know what the situation was in France.) But in Rouen, the choristers were responsible for Sunday vespers and surely used a chant book of some sort. The LU would have served their purposes because it contained the chants for the Office as well as for the Mass. The GR, of course, did not include the Office. What chant book did they use? Is my hypothesis foiled?
In any case, my purpose is not to compare the LU with the GR.
Thanks, GerardH.

In Rouen, 1912-18, my guess is that the change in chant method was simply completing the move from the pre-1908 books to the 1908 graduale and other Roman books published up through 1912 such as the Antiphonale. Rouen had a number of diocesan proper masses, offices and processional liturgies which would also have needed updating. I have no insights into how the differences between the Mocquereau and Pothier rhythmic interpretations were navigated.

With regard to the Liber Usualis being used in Rouen, I think you're right that this was the case. But I don't have evidence for the 1912-18 period. The first Latin edition based on the 1908 graduale was published in 1914, and material updates were included in editions through the early 1920s. Latin editions of the Liber were published prior to 1914, but the chants often didn't match the Roman books.

The Rouen diocese was keen on the Liber to the point that an edition specific to Rouen began to be published at some point with the title, "Paroissien Romain contenant la Messe, l'Office et les Cérémonies pour les Dimanches et les Fêtes a l'Usage du Diocèse de Rouen". The French edition of the Liber Usualis is Paroissien Romain No. 800 (the oldest edition I have is from 1920). The Rouen edition is No. 3800. I have 1956 and 58 editions of this, both of which have sections for Rouen with an imprimater of 1937. As far as I know, Rouen is the only diocese with its own edition.

All of this suggests Rouen used the Liber Usualis and adopted the Mocquereau rhythmic interpretation. Note that the Lyon diocese produced an extensive supplement which did not include the Solesmes rhythmic signs. I've seen versions of the Lyon supplement dated 1921 and 1937. So there could very well be a Rouen supplement to the Liber Usualis dating from the early 1920s.

@5744SHAW I was just thinking I should buy the book, being a Durufle enthusiast. Should I hold out in hope of a forthcoming new edition?

Thank you, GerhardH, for your inquiry. As far as I know, the publisher has no plan for a new edition. Unfortunately, they don't keep me informed as to their plans. They published a paperback edition without telling me.
Jim Frazier

Also, as I read the chapter:

The Solesmes family/other monks (I can’t remember who exactly, but I believe Mocquereau and Pothier were on the same side) had already argued for the restored Sanctus-Benedictus, but Pius X decided otherwise. Dom Combe refers to this in his history of the Solesmes chant restoration (which is of course from Solesmes’s POV, meant to defend the abbey, but it’s still worthwhile), and then it was done, at chant Masses, in 1957’s instruction De Musica Sacra. Obviously, a combined choral Sanctus-Benedictus is impractical, so the larger point remains.

The Fauré usage for the communio eschewed by Duruflé makes sense at a low Mass or at a Mass where there is no communion, since prior to 1960, the deacon would have sung the Confiteor,* or even if it’s merely recited by the MC at a sung Mass without deacon and subdeacon, you still recite the Ecce Agnus Dei and Domine, non sum dignus, and not doing so audibly seems strange to me. So did Duruflé split them musically to allow for communion at Requiem Masses? One would have to specifically investigate this practice in the late 19th and early 20th centuries particularly in France/Paris to get a better answer.

*Stercky, the author of the ceremonial used in many French dioceses and congregations (he was a Holy Ghost Father) writes, in the 1935 edition (the last!) that if the celebrant distributes communion « pour une cause suffisante » then the deacon recites the Confiteor « à voix haute » which is to say recto tono, without the chant melody (in both Latin and in French, recited is a euphemism for this; it doesn’t mean to read in a speaking voice) but the above problem remains the same. The music either has to finish, or you do things that clash with it, which is cacophonous, or you recite things in a speaking voice that the congregation can’t really hear and can barely follow.

In any case, there was not apparently a ban on distributing communion, so in that context, splitting the Agnus from the communio makes sense.