The 1926 D'Oyly Carte Mikado

Cast The Mikado Darrell Fancourt Nanki-Poo Derek Oldham Ko-Ko Henry Lytton Pooh-Bah Leo Sheffield Pish-Tush George Baker Go-To T. Penry Hughes Yum-Yum Elsie Griffin Pitti-Sing Aileen Davies / Doris Hemingway / Beatrice Elburn Peep-Bo Beatrice Elburn Katisha Bertha Lewis Conductor: Harry Norris

Elsie Griffin as Yum-Yum Pearl GEM 137/8

Between 1917 and 1925, HMV made nine acoustical recordings of G&S operas. The invention of electrical recording, which allowed far higher fidelity, prompted HMV and D'Oyly Carte to begin recording the operas afresh, starting with the most-popular opera, The Mikado. The year 1927 was the height of the so-called "golden age" of G&S singing, and it is doubtful that any better cast than this was assembled in the twentieth century. Lytton, Lewis, Fancourt, Oldham and Griffin are all justifiably rated as G&S legends. The one non-D'Oyly Carte singer among the bunch is George Baker, who was included in nearly all the G&S sets of the period because of his excellent diction, even though he never sang with the Company.

Although multiple-casting of roles had once been commonplace, by 1926 it was becoming less frequent. Thus, the allotment of three singers to Pitti-Sing seems curious, to say the least. Michael Walters, who wrote the liner notes for the Pearl re-issue, speculates that Aileen Davies, who sings the bulk of the role, took ill during one of the recording sessions.

Bruce Miller filled in the details of this recording's issue history on the American Victor label:

Victor first used this album design for its issue of the 1917 HMV recording. It continued to use the same cover for the 1926 recording (with the 1917 cast still listed on the included libretto) until the C-12 album, pictured below, supplanted it in 1931.

This new electrical version was issued on black-label (i.e., popular line) records — the only complete G&S opera so issued by Victor. All subsequent complete D'Oyly Carte recordings imported by RCA Victor were issued with the more prestigious Red Seal labels.

RCA Victor persisted in their inconsistency. When they issued a slide automatic version of C-12 in 1934, the labels were still black, despite the red seals they adopted for all the other operas. Two years later, when the 1936 version supplanted older recording, they finally decided the D'Oyly Carte Mikado merited a place among their red seal offerings.

The 1928 "Numerical List of Victor Records" (issued only to dealers) states, on its cover, "Including All Records Announced Prior To [the] January 1928 Supplement." It lists the records among the group of issues from December, 1927. The Victor company apparently substituted it quietly for the previously issued acoustical set. The monthly Victor catalogue supplements do not mention its introduction; unless there was a separately published brochure or other form of advertising (none of which have we seen), there was little fanfare, in the United States, attached to the publishing of this new recording of The Mikado.

Victor Album C-12

Its first appearance in the full Victor catalogue is in the 1928 edition. Here it is not given any album title, album number or record catalogue numbers. The lack of individual record catalogue numbers may be an error of omission, or it may be part of a plan to quietly effect the transition while allowing dealers to sell off any remaining acoustical sets still on their shelves.

The records are identified by side numbers (1 through 22) and artists, confirming that this is the electrical version. Also in the heading is the following statement: "(Recorded in Europe under the direction of Rupert D'Oyly Carte.)" So, someone at Victor must have had a twinge of conscience (or had some correspondence) concerning their earlier advertising puffery with the acoustic version.

The listing for the 1929 catalogue is corrected. It is now listed as "Mikado Album (35860-35870) List Price $13.75", but STILL NO ALBUM NUMBER. No album number is given in the 1930 catalogue, where the listing is identical to the one in 1929.

However, in the 1932 catalogue, the album number C-12 is listed for the first time. To summarize: prior to December, 1927, the available Victor Mikado was the acoustic issue, transferred into the Historical catalogue from the full catalogue [see related article]. After that date, the electrical set gradually supplanted the acoustic one, apparently still using the old "Music Arts Library of Victor Records" album record holder. The electrical "Mikado Album" was finally given an album number in the Concert series, "C-12," sometime during 1931. (This would seem confirmed by comparing it to the other D'OC album numbers and the years they were issued.) Note that there was still no automatic sequence listed.

The slide automatic version, AC-12 (36111-36121), is listed for the first time in RCA Victor's July 1934 monthly supplement, where it is found among the "Other Recent Releases." There was no mention of it in the April, May or June monthly supplements. This curiously unheralded issue echoes the treatment given to the original Victor issue of the set in 1928. The AC version had a short life in America, as it was superseded by the 1936 Mikado in January, 1937.

Bruce supplied a scan of the announcement of this recording, as it appeared on p. 10 of the His Masters Voice monthly New records supplement of May 1st, 1927.

Henry Lytton as Ko-Ko Pearl GEMM CDS 9025

Because it was the first electrical recording, the sound quality on much of this set is less than ideal. But, for the important voices preserved, it is indispensable. It was re-issued on LP and CD by Pearl Records. The Pearl CD re-issue is coupled with the abridged 1933 Sorcerer. I received some email (unsigned, unfortunately) comparing the Pearl and Arabesque re-issues of this recording:

The sound quality is very bad on the Pearl transfer, but that was evidently made from weatherbeaten shellacs. The excerpts that appeared on Arabesque are in much better sound, made, like all of that series, from pristine new pressings that EMI provided. The difference is really astounding and we should all agitate for a complete transfer properly engineered from good pressings. (Note that there is a curious glitch of some sort during Oldham's verse of "The Flowers that Bloom in the Spring" in both transfers. I have come to the conclusion that he swallowed wrong or something during the take and they didn't bother to make another take. Or perhaps there was damage to the master.)

However, Chris Webster disputed this:

The Pearl CD transfer is very crackly but this is certainly not because of weather beaten 78s. If you listen under the crackle to the recording itself the sound is wonderful and bright — there is no sign of wear on the records, to the contrary, these records sound in very good condition. I think Pearl were faced with two possibilities: a) to leave the records as they are, or b) to try to filter out the crackle which would dullen the recording (like some of the non-Arabesque issues). At first the Pearl transer does sound bad, but when I have accustomed my ears to the crackle the transfer is excellent.

Howard Friedman contributed this review of the 78s 2 CD transfer:

I have just received Jim Lockwood's superb transfers of both the 1926 and the 1936 electrical recordings of The Mikado. I have listened to both the original records and also the Pearl and Arabesque reissues of both sets. There are simply no points of comparison. Pavilion Records' (Pearl) penchant for disturbing the original recording as little as possible does not justify the noisy transfers for which they were responsible, and which still seem to be their modus operandi in their current CD listings. While the Arabesque transfers were slightly better, neither of them are in the class of the Lockwood transfers. The clarity of both performances is absolutely outstanding, in fact, quite astounding. It is as though one were hearing both of these performances for the first time.

Although this review is about the sound rather than the performance, it must be said that Henry Lytton's performance far outshines that of Martyn Green, in terms of Gilbertian staging. Bertha Lewis, Sydney Granville, and Darrell Fancourt are, of course, the singers sine qua non in their respective roles of Katisha, Pooh-Bah, and the Mikado. And whoever debunks Josephine Curtis' sole G&S recording has certainly not heard this transfer. While she is no Bertha Lewis, her performance is most creditable and enjoyable.

Jim Lockwood deserves high praise for these outstanding transfers, and for the many others for which he has been responsible.

Issue History

Recording Details

Recording Sessions

Notes:

1. The table above shows the published takes in the order recorded. Where a session produced both unpublished and published takes for the same selection, the unpublished matrix numbers are shown in italics. Where none of a selection's takes from a given session were published, the entire row is shown in italics, and the side number is blank. Where it is not possible to make a likely guess about the number of takes recorded at a session, this is indicated with .... Where the published take was achieved without a make-up session being required, the unpublished takes for the same selection are not shown.

2. It would have been most unusual for more than three takes of a selection to be recorded at the same session (although it did occur—rarely), and indeed one or two was the norm. Therefore, the recording date for s. 4, for which the fifth take was published, may be incorrect. Or, there may have been earlier recording sessions not known to us.

3. For s. 1, 2 & 4, Francis, Francis gives the recording date as 29 September. Thanks to correspondent Peter Chaplin for providing the correct date of 8 November.

4. For s. 3, Francis gives the recording date as 29 September. Ths appears to be an impossibility, as it would place Cc9495-1A among the first matrix numbers assigned to this recording. The date 26 November appears to be more likely, as the preceding matrix numbers in sequence were also recorded on that date.

5. For s. 5, Francis gives the recording date as 19 September. The next recording session occurred some three weeks later, during which time HMV had assigned some 51 matrix numbers to other recordings. This indicates that the side was certainly not made at the first session, as indicated by the recording date given by Francis, but probably in the session of 19 November, as shown. (Francis expresses the date as 19ix26, so the error was as simple as the transposition of i and x).

[Thanks to Howard Friedman for his substantial contribution to these notes. —ed.]