By the first decade of the 20th Century, where we pick up our earliest recorded evidence, Marguerite, Valentine, St. Bris, and Marcel were holding to their designated types. But Urbain had become a contralto, Nevers was assigned to baritones, and Raoul was a newer-style dramatic tenor. In one of our two latterday live recordings (the famous La Scala revival of 1962), that distribution holds, except that the Urbain is now a modern high mezzo-soprano, and the Valentine a dramatic mezzo; in the other (a Vienna concert performance of 1971), Urbain and Valentine are restored to sopranodom, and Raoul undertaken by a tenor who cannot be called heroic, but whose more lyrical instrument has enough strength to aspire to the combination of presence, refinement, and dynamic control that is called for by the writing.

From early in its performance history, Les huguenots was subjected to heavy redaction. Shaw was complaining of this in 1891 (“musical butchery committed a half-century ago”), not because he thought there should be no cuts, but—interestingly, in view of the speculations on the work’s theatrical values entered above—because he thought they should be reconsidered from the dramatic p.o.v. (“In accordance with the taste of that Rossinian period, the whistleable tunes were retained, and the dramatic music sacrificed . . . the score needs to be recut, so as to bring the acting version up to date . . .”) He was not wrong when he called Act V a matter of “rags and tatters.” He may have been referring to redacted remnants, but even if presented in full it is perfunctory and uninspired, and I gather that by the 1890s it was customarily dropped altogether. Yet the story can’t end with Raoul going out the window. The Protestant women must be heard to sing from within the chapel, and then to stop; Valentine must convert; she, Raoul, and Marcel must meet their brutal ends, and St. Bris make his awful discovery. There being no exalted music to carry us through, the act’s effect must depend entirely upon its theatrical representation.

The cuts in our two modern performances are far too numerous to catalogue here. They include entire numbers or verses of numbers (the second verse of Raoul’s Act 1 Romance; Urbain’s Alboni interpolation; the Bohemian Song and Dance near the top of Act 3; Valentine’s Act 4 Romance; the chorus and dance that once opened Act 5, etc.), as well as dozens of internal cuts, short and long, in recitative scenes and ensembles like the finales of  the first three acts. On the whole, I find the La Scala cuts less crippling than the Vienna ones, owing, I’m sure, to the fact that they were made for a fully staged production.

And that production was one of the splashiest of the postwar era, drawing together a collection of singers who could without heavy spin be dubbed “seven stars,” directed by Franco Enriquez in elaborate Romantic-realist sets by Nicola Benois, and conducted by Gianandrea Gavazzeni. It was sung in Italian, as were many (perhaps most) of those outside France in the work’s heyday. If we back off any attempts at Meyerbeerian connoisseurship and simply listen to it as a big bouquet of splendid voices in full bloom, it’s almost always enjoyable, and frequently thrilling. Certainly no such cast could be assembled today. And while I highly value connoisseurship (see below), I think we must remember that creating ineffable effects for the old recording horn is not the same as making a complete role register in the opera house. To zoom in a little closer on the singing of the La Scala performance, taken on its own terms, and considering the female roles first: