In the preface to his Annals of Opera, 1597-1940, Alfred Loewenberg explained his scholarly approach. ‘There are no descriptions of plots’, he wrote, ‘no musical analyses, no personal critical comments. The facts are to speak for themselves, and every care has been taken to verify them so that they may serve as a safe ground on which to build a real history of opera, yet to be written.’ Given that the book goes on to list well over three thousand works in chronological order of their premieres, the full scope of this history might seem hard to imagine, but the information provided for each entry does nevertheless suggest what the priorities of such a project might be. Each piece in the Annals is listed by date, composer and title, together with place of first performance, number of acts and librettist; most then receive a sentence or two of description, sometimes more (and not always as critically neutral as the author’s opening remarks imply).