W. S. Gilbert and Arthur Sullivan. The Yeomen of the Guard. Full Score. Edited by Colin Jagger, with David Russell Hulme. Oxford: Oxford University Press, 2016. [Contents, p. iii; preface, p. v-vii; sources, p. vii-x; editorial method, p. xi-xv; critical commentary (with appendices), p. xvixxxv; dramatis personae & orchestra, p. xxxvi; score, p. 1-407; musical appendices, p. 408-20. ISBN 978-0-19-341313-9. $95.]

W. S. Gilbert and Arthur Sullivan. The Yeomen of the Guard. Vocal Score. Edited by Colin Jagger. Oxford: Oxford University Press, 2016. [Contents, p.v; introd., p. iv; textual notes, p. v-vi; dramatis personae, p. vi; score, p. 1-204; appendices, p. 205-9; index of vocal ranges and dialogue, p. 210. ISBN 978-0-19-338920-5. $23.50.]

W. S. Gilbert and Arthur Sullivan. Iolanthe. Full score. Edited by Gerald Hendrie, with Dinah Barsham, and HelgaJ. Perry. (The Operas, 6.) 3 volumes. New York: The Broude Trust, 2017. [Part A, overture and act I: publisher's pref., p. vii; acknowledgments, p. ix-x; contents, p. xi-xii; editorial policies, p. xv-xix; sigla, p. xxi; dramatis personae & instruments, p. xxiv; score, p. 1-331. Part B, act II: contents, p. vii-viii; editorial policies, p. xi-xv; sigla, p. xxvii; score, p. 1-194. Part C, commentary: contents, p. vii-viii; introduction, p. 1-17; libretto, p. 21-62; critical apparatus, p. 65-157; musical appendices, p. 161-90; literary appendices, p. 193-211; bibliography, p. 215-17. ISBN 0-8540-3006-X. $350 (inclusive of all three parts).]

Looking back, what was the most significant work for the English (or even English-language) musical stage of the nineteenth century? Of the titles that come to my mind, the bulk if not the whole of the short list would be from among the collaborations of William S. Gilbert and Arthur Sullivan. These fourteen works are remarkably varied, although there is an unfortunate tendency not only in popular culture but in music history textbooks to pigeonhole their oeuvre, regarding the pieces as little more than a string of clever words spat out over innocuous accompaniments. The patter songs may be among the most memorable--and certainly the most easily and frequently parodied--aspects of the Savoy operas, but they have only contributed to the too-easy dismissal of the lot.

For more than a century now there has been a steady stream of publications about the G&S canon, but the librettist has been somewhat better served than the composer. It is easier to write about words than about music. Moreover, Gilbert's texts have been more readily available--not only with a number of early authorized editions, but particularly from serious attempts at a scholarly edition. While Reginald Allen's The First Night Gilbert and Sullivan: Containing Complete Librettos of the Fourteen Operas, Exactly as Presented at Their Premiere Performances, together with Facsimiles of the First-Night Programmes (New York: Heritage Press, 1958) is not quite what it claims to be, it at least made the enthusiast aware of variants in the quasi-sacred texts. Allen tried to establish the text as originally performed, relying with too much confidence on the reading of the libretto distributed (as far as he could determine) at the premiere of each show. Starting in the 1980s, Ian Bradley employed a much more thoroughgoing approach, appearing most recently in his The Complete Annotated Gilbert & Sullivan (20th Anniversary Edition [New York: Oxford University Press, 2016]). Bradley looked well beyond the published librettos and subsequent authorized editions, scrutinizing the preproduction copies submitted for license from the Lord Chamberlain's office, as well as the early promptbooks and Gilbert's sketches and drafts extant now at the British Library. Even so, Bradley did not consult musical sources, and thus has no substantiation for annotations such as "In early performances Elsie had a longer solo here" (p. 922).

Far surpassing Bradley in a comprehensive attempt at textual authority is The Variorum Gilbert & Sullivan, edited by Marc Shepherd and Michael Walters. To date, only one volume has appeared (New York: Oakapple Press, 2015), containing the first four operas. Although their edition presents only the verbal text, Shepherd and Walters draw upon the full range of sources available--manuscripts (Sullivan's as well as Gilbert's, letters as well as lyrics), the license copies, the various published librettos and scores. This difference is significant, as the musical sources contain vital evidence for understanding what was or was not performed--what was set and scrapped, and what was never set at all.

So much for the words, but what of the music? I am not the first to bemoan the paucity of credible editions of Sullivan's half of this collaboration. At least as far back as 1928, Thomas Dunhill emphasized this point with patriotic incredulity:

Owing to the fact that the full scores are unavailable, people are obliged to form their estimate of the music from the published vocal scores with the accompaniments transcribed for piano. These are not ill adapted for the practical purpose of rehearsing, but they are rather clumsily arranged, and give an incomplete idea of the composer's original conceptions.... It seems incredible that the British people should be denied access to the work of one of their greatest musicians in its only proper and authentic form. (Sullivan's Comic Operas: A Critical Appreciation [London: Edward Arnold, 1928], 216-17.)

This is not the place for recounting the gradual change in this situation. A good summary has been given by David Russell Hulme ("Adventures in Musical Detection: Scholarship, Editions, Productions and the Future of the Savoy Operas," in The Cambridge Companion to Gilbert and Sullivan, ed. David Eden and Meinhard Saremba [Cambridge University Press, 2009], 231-42). Indeed, it was Hulme's dissertation ("The Operettas of Sir Arthur Sullivan: A Study of Available Autograph Full Sores" [PhD diss., University of Wales, 1986]) that established the field of Sullivan source studies, and his own edition of Ruddigore (Oxford University Press, 2000) is clearly the model for Colin Jagger's edition of The Yeomen of the Guard, for which Hulme also served as "consultant editor." (Hereinafter, Jagger's full score will be cited as "FS"; the vocal score as "VS.")

Jagger's Yeomen is very accurately described as "a scholarly performing edition" (VS, p. iv). Jagger has collated sources relevant to the work from its origins through the D'Oyly Carte revisions of the 1920s, seeking to present a text as authorized by its creators--not necessarily that of the first night (already before the premiere Gilbert and Sullivan had agreed on cuts to be made immediately after the first night), but in the settled state of its original run. He has nonetheless borrowed clarifications from later versions of the text--particularly stage directions indicating characters' entrances and exits, some of which remained unspecified in print even after Gilbert's supervision of not only the 1887 premiere but also revivals in 1897, 1906, and 1909. (One of these even now seems curiously to function more as an observation than a direction: "In the meantime, the Chorus have [sic] entered" [FS, p. 301; VS, p. 145].)

The back cover of both the FS and VS announce that this edition "returns to what was performed during the original Savoy Theatre run," but the inclusion of material that never once was performed in that run suggests that the editor is advocating yet another version. There are essentially five portions of discarded material--and, of course, each has its own story. Most straight-forward of these is the original version of Fairfax's first song, "Is life a boon?" (no. 5 in act 1). This version made it into rehearsals, but was discarded in favor of a new--now standard--version just days before the premiere. In Jagger's edition, the original setting appears as an appendix in the back of both FS and VS, although a note in the FS indicates that in the orchestral parts both versions of this song are in situ--indeed, the printed sequence of the two versions of no. 5 in the individual parts varies "according to the ease of page turns" (FS, p. 104). As both versions are given the same item number, some confusion seems inevitable in rehearsal.

This situation set me wondering whether there would ever be a production that would opt entirely for the original version over the one that superseded it. I doubt it. Although Fairfax is a rather despicable character, and there's something of an appropriate swagger to the Allegro pesante 6/8 of the original (indeed, with a touch of Verdian brindisi), somebody in 1887 clearly recognized that the original setting did not do justice to Gilbert's existential lyric--the lyric that the librettist would later choose to be inscribed on the memorial monument for Sullivan on the Victoria Embankment. Sullivan's second setting (Andante espress., partially reworking ideas from the first setting) is justly celebrated. A production of Yeomen without it? It would be like a production of Carmen that replaced the famous Habanera with Bizet's first attempt, "L'amour est un enfant de boheme!" In either case I can only imagine a performance in which the original is immediately followed by an exclamation equivalent to "O Freunde, nicht diese Tone!" before launching into the familiar version.

Desirable as it is to have these extra numbers available for study, the layout of the edition--placing most of this material in its performing sequence rather than as an appendix--implies that the creators regarded these as performance options in that original run. In fact, each was explicitly cut, either because of a deficiency in the number or because of a weakness in the overall pacing. In one instance, this layout is a matter of convenience: couplets for the Third and Fourth Yeomen near the beginning of the act 1 finale were cut just before the curtain went up by simply eliminating a repeat. It is much more practical to include this material in situ (as Jagger does) with a repeat than to print it separately at the end of the volume--and this cut is noted in a footnote (p. 193). Similarly, Meryll's song "A laughing boy but yesterday" (no. 3a in act 1) was a cut decided immediately before the premiere, but in fact it was sung that evening (only), apparently as a courtesy to the D'Oyly Carte veteran bass-baritone Richard Temple, as it was his only solo number in the piece. Why Jagger does not relocate this number to an appendix is not clear to me, unless he is hoping it will be performed more frequently henceforth.

Wilfrid's song "When jealous torments rack my soul" (no. la in act 1) was never once sung in the original run, but rather cut about two weeks before the premiere. Indeed, it was so thoroughly expunged that Jagger has had to piece together materials from the original New York production--materials prepared and dispatched overseas before the cut was made. Even so, the textual complications here are fascinating, with an independent horn part (apparently the sole survivor of yet another version of the number) dated fully a month after the American premiere. How this part came into existence remains a mystery, and Jagger's commentary on this curiosity makes compelling reading for anyone intrigued by textual complexity (FS, pp. xxix-xxxi). Like Meryll's "A laughing boy," this song is the only solo number for its singer, but that is the only point in its favor--and not a point that seems to have bothered its creators. W. H. Denny, who created the role of Wilfrid, was new to the company; there was no particular motive to reward him with his own solo moment. Out it went.

The fifth discarded passage is the original ending of the penultimate number of the opera, the duet "Rapture, rapture!" Gilbert originally placed this before the announcement of Fairfax's reprieve, so that the duet was followed by a final passage of dialogue before the finale. Sullivan's original ending was overlong, considering that the characters still need to be onstage (seen "going off") when Phoebe was to speak the next line of dialogue. His solution was to truncate the coda. When the collaborators moved the ensuing dialogue to before the duet, Sullivan expanded the ending again, but this time coming to a half cadence to segue into the finale. Curiously, Jagger includes the original longer ending (immediately after the revised ending in FS, in an appendix in VS) while not facilitating the performance of the scene that requires it: the original version of the dialogue is given as appendix E to the critical commentary in FS (p. xxxiii), but is absent altogether from VS--the source that the actors who would speak the lines would be using. Jagger errs, I think, with his editorial suggestion "Attacca Finale" with the two earlier versions of the ending, as these were not intended to lead directly into the finale at all. In FS the instruction rightly appears in brackets as an editorial emendation; in VS it is without brackets, suggesting that it is somehow authoritative.

Jagger restores the famous duet "I have a song to sing, O!" (no. 7) to its original key (D major). Indeed, he makes no provision for it to be performed in E-flat (as has been standard since at least the 1897 revival). He does not dispute the composer's authorization for this change: "we are unlikely ever to be sure what Sull.'s reasons were, but his original intention is clear, and it is almost certain that the duet was in D during the first production" (FS, p. xxi). He argues that "since the duet segues from the previous number which sets us up for D, any other key would seem incongruous (the music is reprised in the Finale of Act II, where again it is in D)" (pp. xx-xxi). The last portion of the previous number is in G Lydian, but it is not at all clear how the segue is to happen, as there is an extended dialogue over music. It has no conclusion, but has a repeat marked. Was the orchestra intended to cut directly to the duet whenever the dialogue finished, from whatever point in the playout (treating it, that is, as a "vamp till ready")? If so, the fact the last harmony Sullivan writes is A major--the dominant of the new key--is immaterial. E-flat is certainly an arresting change, though incongruous seems too strong a term, especially given the composer's authorization. Jagger's edition generally enables options, but not this one: users are forced into Sullivan's original key. I think this is a mistake: the change to E-flat may well have been prompted by aesthetic considerations that made an "incongruous" key worthwhile. I speculate: for this strolling minstrel number, the ostinato drone (a perfect fifth sustained in bass and cello, but with sustained chords above) accompaniment suggests a hurdy-gurdy. This effect works well on paper, but might it have been found too harsh in performance? Moving it up a semitone necessarily softens the string timbre by eliminating even the possibility of sustained open strings. Jagger may prefer it in D, of course, but it is editorial overreach to discard Sullivan's second thoughts--as we cannot know what motives he might have had for this change.

The organization of FS is puzzling. There are two distinct sets of appendices: at the front of the volume, following hard upon the critical commentary (and before the musical text itself), are appendices A-E dealing mainly with verbal text. At the back of the volume is the score of the original version of "Is life a boon?"--labeled "Appendix 1." There is no appendix 2 in FS, but Jagger's appendix E (the original final dialogue, discussed above) would fit better here at the end, as an alternate performance text. Appendices A-D are discussions of textual matters and belong where they are, although it would be helpful to have them listed in the table of contents. More puzzling is that VS does have an appendix 2--the original ending of "Rapture, rapture," which in FS is folded in alongside the main text. VS lacks the critical commentary (and thus also appendices A-E), but has a very useful summary of "the most significant points pertinent to the vocal score" (p. v). These will no doubt be useful in forestalling seasoned Savoyards' quibbles about some of the original readings that Jagger has restored. (Another difference in VS is an "Index of Vocal Ranges and Dialogue," which will certainly be useful in planning rehearsal schedules.)

All this notwithstanding, this is a fine edition. Jagger has (to borrow a phrase from Ralph Vaughan Williams) "washed its face" very thoroughly, and the work is presented admirably. Original details that unaccountably went by the wayside have been restored--particularly in ensembles when different characters should have different lyrics. It is clearly designed with performance in mind, and it should give a solid textual foundation for new productions of the opera willing to invest in it, as well as becoming the default musical source for those of us studying the work.

The other edition under review here is the Broude Trust's Iolanthe, in three bulky volumes totaling more than 750 pages. The publisher's preface for the series describes it as "a pragmatic compromise between the scholarly and the practical" (part A, p. vii), and corresponding performance materials are promised. These tomes represent decades of work credited to three distinct scholars: the late Dinah Barsham, her husband Gerald Hendrie (who ultimately completed the project), and contributor Helga J. Perry. At the annual meetings of the American Musicological Society, I have for the past two decades stopped by the Broude booth to ask about the progress of their Gilbert & Sullivan: The Operas edition. Ronald Broude's reply that "Iolanthe is in first proof" or "second proof" (and eventually "third proof') became a regular refrain of these meetings for me--but it is a testimony of the painstaking dedication of the publishers to such a costly project. Gilbert and Sullivan may be long overdue for a serious scholarly edition, but the Broude Trust is intent on seeing that they get one.

If Oxford's Yeomen is fine, Broude's Iolanthe is superb, and well worth the decades of waiting. It is--as with the others in the Broude series--beautifully presented. The number of typographical issues I detected was vanishingly small (e.g., a missing rest here, some braces smushed into brackets there), and the number of actual and suspected errors minute. For Broude's Trial by Jury and HMS Pinafore, errata lists for the full scores were included when the subsequent vocal scores were published, and I will communicate my brief list to the publisher rather than belabor a reader here.

Broude tends toward a "one-stop shop" approach in this series: a comprehensive introduction to the work and its genesis; a critical edition of the libretto (including two sets of footnotes: one for variant readings, and a second for annotations to elucidate potentially obscure references); a critical edition of the full score (with discarded numbers), with separate critical apparatus; and supplementary materials including some of Gilbert's 'Bab' Ballads that served as precursors to elements of the plot of the opera, and--scattered throughout--his comic drawings illustrating characters at various key moments; for Iolanthe a fascinating extra is a collation of various arrangements of the beginning of the "March of the Peers" in act 1 for onstage brass band, illustrating a variety of practices in the early years of the D'Oyly Carte productions. If the Oxford Yeomen becomes by default the standard edition of a musical text of the work, the Broude project intends to establish standard editions for the text tout court, not just the musical text. One significant absence here is Gilbert's preliminary plot sketches and drafts for Iolanthe preserved at the Pierpont Morgan Library, and not yet published in any widely-available form; even considering the space it would consume, this edition would have been an obvious place to publish this (at least until the relevant Variorum volume appears), so it strikes me as an unfortunate omission--particularly as the editors cover it in a mere four paragraphs. A lengthy description, quoting extended segments, is included in Andrew Crowther, Gilbert of Gilbert & Sullivan: His Life and Character (Gloucestershire: History Press, 2011), 156--68. Crowther's work is not cited in the Broude volume, and indeed the Broude bibliography is not only limited but shockingly dated: of the thirty items of "secondary literature and reviews of modern productions" (part C, p. 216), only seven date from the present century (and two of these are tangential in the extreme). Compare this with twelve items published in the period 1970-1990--the period, that is, when the work on this project began--despite a significant increase in studies concerning the Savoy operas in recent years. My sense is that the introduction is dated in a way that the rest of this edition is not.

One other item that would have been extremely valuable to users of the edition is a letter from Sullivan in late October 1882 to Alfred Cellier, who would be conducting the New York premiere, transmitting specific instructions about a number of details in the score. Hendrie cites this letter frequently in the critical commentary of the musical text, but the full text would be useful. Some of these instructions are truly surprising, as for example Sullivan's remark about Private Willis's "When all night long," which opens act 2: "1st eight bars of song to be sung ad lib: like a gentleman at a public dinner without accompt" (part C, p. 132). Hendrie adds a comment that "there is no evidence in extant musical sources that the song was actually performed this way." Nonetheless, I think that this oddity merits a footnote on the relevant page in the score--although it doesn't get one. Hendrie does include a footnote for Sullivan's instruction that the last phrase of the Lord Chancellor's song "When I went to the bar as a very young man" should be "spoken through accomp," indicating "I have taught [George] Grossmith to sing it thus" (cf. part C, p. 119, and part A, p. 212). There is also a footnote regarding the choral echo ("I wonder") at the end of "Oh, foolish fay"--something Sullivan specifies in the letter, but which remains absent from the published scores until the 1920s. Hendrie concedes "it was apparently an early performance tradition" (part B, p. 53)--indeed, an instruction from Sullivan before the first night. Such situations prompted me to wonder how important it was to Sullivan that his published vocal scores aligned with the details of his intentions on stage. Gilbert took pains to get his text the way he wanted it in print, both in librettos for distribution at the theater and subsequently in published collections of his plays. By contrast, the small revisions to the first English edition of the vocal score seem not to have been published until sometime after Sullivan's death on 22 November 1900. (After the initial London production closed on 1 January 1884, the D'Oyly Carte company did not revive Iolanthe at the Savoy Theatre until 7 December 1901. That revival might have been the impetus to make some revisions to the vocal score, but the Broude team could not locate a copy of the revised first edition that could be dated before 1911.) D'Oyly Carte licensed amateur productions, and for these the published vocal scores would easily suffice while not capturing every performance detail that the professional companies (i.e., D'Oyly Carte's main London company and his several provincial touring companies) would give. Much has been made of Gilbert's scrupulous preparation of prompt books, and insistence of faithful adherence to the letter, but there is no evidence to show that Sullivan cared at all that the same fidelity be given to an accurate musical text.

The main textual challenge faced by Jagger, Hendrie (et al.), and indeed any editor of the Gilbert and Sullivan works (or works for the musical stage generally) is the lack of a single authorized text for the whole work. Both the Oxford Yeomen and the Broude Iolanthe exhibit an editorial principle of "divided authority": a certain state of the published libretto is the authoritative source for the dialogue; a certain state of the published vocal score serves likewise for the vocal parts, Gilbert's lyrics (which Sullivan sometimes altered), and word setting; and for the instrumental lines, the ultimate authority is Sullivan's composing score, faute de mieux (as Hendrie acknowledges, given the sketchy and at times skeletal state of Sullivan's autographs). This complex editorial situation is inevitable, and of course there are very many situations in which two or more of these principal authorities conflict, to say nothing of the many sources further down the textual chain. The editor's task is not one of merely combining the authoritative elements of these three sources to produce a single document: that could be done mechanically and mindlessly. Far beyond this, the editor must interfere with the readings of these principal authorities to produce what--as best can be determined--a single authorized edition would have looked like, had it only been produced. For example, in the composing score, Sullivan sets an outburst from the chorus of Peers with this text:

O lucky little lady!

Her Strephon's lot is shady;

His rank, it seems it vital,

And Countess is the title,

But of what I'm not aware!

(Iolanthe, act 1 finale)

Anyone familiar with Iolanthe will recognize immediately that the first four lines have an extra initial syllable--a syllable that may well be the composer's addition to the text, as this version is not found in any of Gilbert's librettos. Indeed, these extra syllables do not survive even into the first edition vocal scores (English or American), and are extant only in Sullivan's autograph. Hendrie has selected the first edition English vocal score in its second state as his authority for the vocal lines and text setting, so these initial syllables are relegated to the critical commentary. But as the full orchestra (minus the scurrying violins) is doubling the chorus here, Hendrie emends the orchestral parts by deleting the initial eighth-note for each phrase to match the rhythm of the vocal score (part A, pp. 245f.; cf. part C, p. 125). The alternative--blindly following the prescribed division of authority--would have yielded different rhythms between chorus and orchestra precisely where Sullivan set unison rhythms. (Although early orchestral performing materials survive and were consulted, the only known material extant from the original production is a violin I part, and thus cannot answer the question of what rhythm was played at the Savoy in 1882.)

The last of these lines--"But of what I'm not aware!"--presents another problem. In some vocal scores--and, indeed, in a revised text in the composing score itself--this line is preceded by a sort of repetition of the one before it: in later vocal scores it reads "Yes, Countess, Countess, the title, the title," while in the autograph it reads "Yes, Countess, Countess will be the title!" Neither of these appears in Hendrie's score (nor even in a footnote), which instead adheres to his specified earlier state of the vocal score, and giving instead "But of what I'm not aware, I'm not aware!" followed by the line that is in all sources, "But of what I'm not aware." It is a small point, but as this same line is then repeated again (with different music) to wrap up this section of the finale, a footnote in the score would have facilitated a choice for performers who might opt (with the autograph and with later vocal scores) to state two lines twice rather than one three times in succession; which, however, is not to say that economy of delivery is a particular hallmark of the Gilbert and Sullivan operas.

To the extent that the Broude series seeks to wed the scholarly and the practical, it is my assumption that what appears on the pages of score is what performers are expected to use, while what appears in the critical commentary is aimed at those pursuing a more scholarly approach. (Will these scores one day be made available as cheaper paperback offprints without the critical commentary?) If I am annoyed at times by the absence of a footnote in the score when I feel a performer's attention ought to be drawn to something, I find myself also bemused to find the occasional footnote that flags something no one would ever dream of performing anyway (e.g., part A, p. 176). Greater care might have been given to precisely what should appear in these footnotes--and this is one way in which the Oxford Yeomen is so much more clearly a practical edition. When Hendrie stoops to practicalities, the effect can be jarring. "The editor suggests the following cadenza," he writes, supplying a short flourish of his own for Phyllis where Sullivan specifies only "Long cadenza" (part A, p. 222). As this is a task that a singer, conductor, or repetiteur should be able to do without difficulty, this editorial condescension is a blot. The same may be said of the annotations to the libretto, which are sometimes quaintly expressed (noting that St. James's Park was "in Gilbert's day known as a venue frequented by women of easy virtue," as if the reader should not be compelled to see the word prostitutes (part C, p. 36), or bizarre choices (as if "affadavit" is a word with some ambiguous meaning or no longer in common usage [part C, p. 34]). The numerous references to "long-standing performance tradition" seem out of place in an edition that is seeking to establish the settled text of the initial run; a reference to "modern performance tradition" (part B, p. 85) even more so.

In one instance I think Hendrie's musical judgment has failed him. Near the end of the act 2 trio "If you go in" (at the words "It's Love that makes the world go round!"--on the word "Love") he argues that an unclear note-head in cornet I should be read as g' (concert e') rather than a' (concert f# as in a copyist's score and the early printed orchestral parts--which essentially changes the harmony from ii (6) to IV in E major (part B, p. 127). The note in the critical commentary reads "Although Afutograph] is not completely clear and the added sixth is not impossible, it seems more likely that Sullivan intended a simple A-major chord here" (part C, p. 142). Hendrie offers no further justification, although he acknowledges that there is a parallel passage (twice) when this music is reprised in the finale ultimo--where not only cornet I but the tenors sustain the concert f#' (part B, pp. 179 and 191). Indeed, the f#' is not at all an "added sixth" but harmonically essential: the e" that appears above it is a melodic dissonance--and indeed not sustained in any of the accompanying parts in the more fully scored passage in the finale. Hendrie's misreading of this minute detail should be corrected.

If I seem overly harsh on Broude's Iolanthe, it is because it will not be bettered anytime soon. If the Broude Trust ever gets around to The Yeomen of the Guard, there is still work to be done, but Oxford's Yeomen will suffice for the foreseeable future. The two editions reviewed here must inevitably be regarded as momentous, nothing short of revelatory in what they open up to the new scholarship of the Savoy operas. These two operas are among the collaborators' best works, and they finally are published in a form that will allow proper consideration of their merits and liabilities. Both belong in the libraries of any program with any interest in the musical stage. Oxford has managed to keep the price of their Yeomen at least plausible for companies that might seek to put on the show (considering that inferior performance materials are available for free download from IMSLP). Broude's series will always be a stretch on the pocketbook, but it truly offers outstanding value for the money.

James Brooks Kuykendall

University of Mary Washington