Notes on Matrix and Catalog Numbers

The first 406 recordings of the Gramophone Company made in London between August 8 and October 31, 1898 in the "original" series were identified by catalog and serial numbers. The former were given prefixes indicating the language or country of origin of the selection, thus, E for English (318), F for French (11), Gr for Greek (1), Gy for German (75), and J for Japanese (1). Before November 1, 1898 the serial numbers, which are not always discernible, began each day with 1, and from that date on the tablets were numbered consecutively from day to day. The only way that one can distinguish identical serial numbers is by the recording date, which, fortunately, was also etched in the central area of the wax-coated zinc recording plate. Needless to say, all of the issued discs were Berliners without paper labels. Not all of the recordings made during this initial period have been identified.

The Gramophone Company used two numerical systems for identifying the recordings. The matrix or serial number indicated not only the sequence of the recordings, but also the recording or processing date, the artist, selection, and often the location. Each engineer filed a Weekly Return every Saturday, containing the names of the artists and their selections, together with the assigned matrix numbers, which were also inscribed on the recorded tablets. These were sent to Hanover, and later to other processing plants, together with the tablets recorded during the previous week. At Hanover (and later other plants) these data were entered into a consolidated Register, together with the catalog number that was assigned to each recording at the factory. Each Register contained information regarding the recordings according to the relevant region or language. The Company’s areas of responsibility were divided into ten regions, largely according to language. Each region was assigned a block of 10,000 catalog numbers, and a set of one or more Registers as required. England was assigned numbers 1 through 9999, Russia 20000 through 29999, France 30000 through 39999, Italy 50000 through 59999, and so forth. When paper labels were introduced, 7-inch recordings were assigned the appropriate number, 10-inch recordings were given numbers with the prefix G.C.-, and 12-inch records received the appropriate catalog number preceded by a zero 0. When the first block of numbers for a particular region or record category was exhausted, a second series was begun with the prefix 2-, a third series with 3-, and so forth, as required. See the various catalogs of Alan Kelly for further details.

The brilliant researches of Peter Adamson (see the Bibliography) make it reasonably certain that Fred Gaisberg spent most of the week from Tuesday, August 2 (Monday, August 1, was a Bank Holiday, when most shops were closed) through Saturday, August 6, 1898 buying and assembling the various chemicals and other materials that he would need for his recording process. The Gramophone Company of London began to make recordings as early as Monday, August 8, 1898. From that date to about November 1901, the issued records had no paper labels, and were identified by the recording engineer or his assistant, who inscribed the date and other pertinent information in the central area of each newly recorded matrix or plate. These details, together with the company designation, catalog number, and later the trademark added afterward, then appeared on every pressing made from the processed matrix (see below).

One of the "myths" of G&T discography is set forth and elaborated upon by Fred Gaisberg himself, in his autobiography as well as in various interviews which he made several decades after the "events" in question, by which time he was the only one left to "tell the tale!" Fred said at those times that the first recording he made in the London Studio in Maiden Lane was of a young barmaid with a reasonably good voice, named Syria Lamonte. This young lady actually made her first recording on September 1, 1898, rather than on August 8, 1898, the latter is generally regarded as having been the first day of the London recordings. Miss Lamonte’s third recording, made on September 2, 1898, is shown on the right below, and bears the catalog number E3005. Alan Kelly and Peter Adamson (see the Bibliography) have both shown that six recordings were made on wax-coated zinc plates on Monday, August 8, 1898. The last of these, E6002, is shown on the left below.

     

It was assumed for many years that Miss Lamonte made the first London recording on August 2, 1898. [Author’s note: Perkins and Kelly were still accepting this date for the first London recordings as late as 1982. See the Bibliography.] This may have been due partially to Fred’s "remembrance" of her recording as having been the first one made in the studio on Maiden Lane, which actually did take place on that date and in that place, but with a different performer, as well as to the fact that the date seen on her issued record is 8-2-98. Even more remarkable is the fact that in his autobiography, The Music Goes Round, Fred states "I actually found my first artist here in the person of Leopold Jacobs." In fact, Jacobs made the first known 10-inch all-wax recording in April 1901, but his first name was not Leopold ─ it was Jacques, although only J. Jacobs appears on his issued discs. Jacobs was the leader of the Trocadero Orchestra which made its first recording, Berliner Gy512, on September 6, 1898.

A quick perusal of the figure on the right above seems to indicate that this is the date on the disc. However, Adamson’s investigations have shown that Fred Gaisberg and Joseph Sanders were at the Cecil Hotel in the Strand in London on Sunday, July 31. As August 1 was a Bank Holiday in England, most shops would have been closed. Adamson has shown that Fred spent most of the following week purchasing the various supplies that he would need to set up the recording studio, viz., sheet zinc, shears, alcohol, various glass and tinned containers, copper wire, and other materials necessary for the recording process. These purchases occupied most of Fred’s time through Sunday, August 7. By that time he was more or less ready to start work on Monday morning, August 8, 1898. Thus, if one scans the disc more closely, it appears that the inscribed date was simply an error in transcription. This is particularly plausible, since the handwritings on the two discs shown below are quite different, indicating that one of them was probably entered by Fred and the other by his assistant, whoever that might have been.

The recordings of the "original" series were assigned catalog numbers according to a system, doubtless instituted by Fred Gaisberg and derived from that used in America and Canada for several years previously, which would survive until the end of the acoustical recording era. This internal system involved the last four digits of each catalog number, the first or first two or even three being used to identify the region or language of the recording, as well as the record size. The fourth number from the right identified the content of the recording, e.g., talking, band, vocal, instrumental, or other, while the fifth number from the right identified the language or region, as outlined above. The details are found in numerous sources, particularly in Alan Kelly’s several catalogs. The American and London categories are outlined in the table below. These categories were used for assigning the catalog numbers to the "original," as well as to all subsequent series. Syria Lamonte’s recordings all received catalog numbers in the 3000 series, indicating female solo vocal, and as almost all of the recordings preceding hers were clarinet solos, they received catalog numbers in the 6000 series. When the base number was less than 1000, the initial 0 was omitted.

On September 1, 1898, a second "unlettered" series was begun, with serial numbers assigned in a continuous sequence daily. This is the major difference between the "original" and "unlettered" series. It must be emphasized that these were serial numbers, not matrix numbers. The catalog numbers assigned to these early recordings were based on the system already established in America comprising some forty or more categories of recordings, which would be used by the Gramophone Company throughout the acoustical recording era, as stated above. This rather fragmented division would be reduced to a more manageable and quite different twenty-eight categories across the Atlantic, as seen in the table below. Some variations were made to accommodate regional differences.

The company’s only official recording engineer between August 8, 1898 and May 1899 was Fred Gaisberg. William Sinkler Darby had been sent to Russia in the latter month by Emile Berliner, more or less secretly, i.e., without the knowledge of either the London office or the Hanover plant. The recordings that he sent back to the Hanover factory were all assigned catalogs in the 20000 series. When they embarked together on the first foreign tour, Fred identified his work with a G added above the serial numbers entered on the record surface. Darby started with an A suffix, which was changed to B, x, and y, for 7-, 10- and 12-inch all-wax recordings respectively, following the tour. Franz Hampe, the first engineer of European origin, began his activities in Munich in July 1902, using C, z, and Hp.

From Alan Kelly’s Holland Catalogue:

"While using zinc blanks, Darby had lettered his matrixes in a numerical series carrying the suffix letter A (upper case) to distinguish them from Fred Gaisberg's work. With the appearance of solid wax blanks, new matrixes were numbered in a new series, this time carrying the suffix B (upper case). The introduction of ten-inch wax blanks caused Darby to begin another numerical series, this time suffixed x (lower case, Roman numeral for ten). As time went on and other experts were employed to make records, the system of suffix letters, properly known as indicating letters and used to show who had made the particular recording, became quite complicated."

As additional engineers were added to the Company’s staff, and as new record sizes were introduced, it became clear that a new system of serial numbers was required. Will Gaisberg, Fred’s brother, is generally credited with having devised it. This simple system, which was probably introduced early in 1904, consisted of assigning a group of three letters, for 7, 10, and 12-inch recordings, to each engineer, to be used as suffixes to his serial numbers. Fred was allocated a/b/c, Will was given d/e/f, and Darby had g/h/i or j, while Franz Hampe was given k/l/m. Each recording engineer could then travel to his assigned locations and use the next serial numbers in his block, rather than having to use a block of numbers assigned by the London Head Office. Then any master recording, metal part, or finished record could be identify as to the recording engineer and the record size.

Before the system was implemented throughout the staff, Will made a series of recordings in all three sizes in Milan in October 1903, using the serial numbers CON 20 to CON 99 for 7-inch, CON 100 to CON 499 for 10-inch, and CON 500 upwards for 12-inch recordings, respectively. It was not until it was discovered that Will had made a series of recordings for the Zonophone label using the prefix RAD, that it became apparent that he was using his middle name CONRAD! Before the letter triplet system was in use, secondary stampers were embossed with only the serial number in the runoff area. After the system was introduced, new stampers of older recordings generally had the matrix number together with the recording engineer’s appropriate suffix.

      

Note that the matrix suffix on the disc at the left above is a rather large capital Z, rather than the lower case z expected here. These are typical of Hampe’s recordings. The disc shown on the right was recorded by Will Gaisberg in Milan in October 1903, and was assigned the matrix number CON 187. Close examination reveals that the original matrix number was removed and replaced by an embossed 187─rather unusual for a first stamper pressing.

Following the introduction of the letter triplet system, each of the four engineers continued his respective numbering blocks for each disc size, while adding his assigned suffix to each respective series.. Will Gaisberg made his first recordings using matrix numbers with the new suffixes in Stockholm in February 1904, following his CON 142 with 144d, CON 402 with 403e, and CON 750 with 751f, respectively. These included about 91 7-inch, 79 10-inch and 13 12-inch recordings, and were followed by a similar group made in Copenhagen in March 1904. Alan Kelly indicates that Fred Gaisberg may have used his new suffixes as early as December 31, 1903. Darby seems to have used his assigned suffixes on recordings made in Berlin in mid-1904, while Franz Hampe may have used his new suffixes a month later, probably either in Riga or Tallinn, Estonia.

[Author’s note: it may be of some import to observe that of 31 of Caruso’s G&T discs pressed from 20 different stampers and showing CO. markings, none show the b suffix customarily added to earlier recordings attributed to Fred Gaisberg. The b suffix is seen on nine of these issues after July 29, 1905 when the use of the CO. marking was discontinued. This is equally true for similarly marked Tamagno issues. Further, all uses of the b suffix on Caruso issues and the e suffix on Tamagno issues are found only on records issued after that date. If this actually ceased after July 1905, then the institution of the triplet letter suffix system may have been later than previously thought. On the other hand, there is no reason to assume that the various suffixes were entered on older recordings immediately following the institution of the new matrix system. Surely the several recording engineers working at that time would have been the first to be informed of the new system, and probably the processing plants would have been the last. Although suffixes d, e, and f are listed as having been used in February 1904 on recordings made by Will Gaisberg in Stockholm, it is not certain that the observed discs were pressed from first stampers. However, Liliedahl specifies that d, e, and f suffixes appear on recordings made in February 1904, and B, x, and y, Darby’s earlier letters, on recordings of the previous session in late 1903.]

Despite the simplicity of the new system, at least three major types of errors are known to have occurred, the first of which concerns the recordings made by the two engineers, Cleveland Walcutt and Charles Scheuplein, assigned to the Paris branch. For details, see FRENCH RECORDINGS below. The second type consists simply of the same matrix or serial number being used more than once, sometimes as many as three times, which happened rather frequently. One of the most notable of these errors (and there were plenty!) was the assignment of the serial number used for the last recording by the soprano Amelia Pinto on the morning of Friday, April 11, 1902, to the first recording made by Enrico Caruso in the same recording room in Milan that afternoon! (Jerrold Moore, Fred’s erstwhile biographer, says that Fred, in his excitement, entered the wrong matrix number himself, but this is doubtful. In actual fact, the error was probably made by Fred’s assistant, who was probably his brother Will, one of whose duties was to inscribe the next matrix number after the recording had been made!) Amelia Pinto’s final recording has the matrix number 1782-BG, which can be seen plainly impressed through the label of a first stamper pressing, on the left below. Caruso’s first recording was assigned the same number, which appears impressed through the label as 1782-nB, as seen on the right below.

 

 

Although the numbers themselves appear to be in the same handwriting, it is extremely doubtful that either Fred or his assistant entered the suffixes; they were probably entered at Hanover in order to distinguish the two recordings. From several copies of the Pinto recording that I have seen, it seems quite likely that every effort was made to remove the duplicate serial number 1782 and substitute the one assigned to her first take of the same aria, Vissi d’Arte, which was 1780. Several stamper II pressings with catalog number 53234X have been seen overprinted with 53232X, but with no matrix numbers. A stamper IIII pressing with catalog number 53232 X bears the matrix number 1780, that of the first take. The two takes can probably be distinguished best by the fact that the first, matrix 1780, is 3 minutes and 5 seconds long at 71.29 revolutions per minute, while the second, matrix 1782, is 2 minutes 58 seconds long at the same speed, which Aida Favia-Artsay had determined for the Caruso recordings made on the same afternoon. Moreover, the first take begins with some background chatter, which was probably the reason for the second take. (See below under Celebrity Labels)

As a further note, stamper I pressings of the Caruso 1782 show a superscript a following the matrix number, which is not seen on stamper II pressings. One might conjecture that, when the Weekly Return was being filled out it was noted that two recordings had been given the same matrix number! It would therefore seem logical that one of the recording tablets should be marked in some manner to distinguish it from the other. Hence the addition of the a. However, the two recordings were already distinguished from one another by the suffixes BG and nB. What is even more inexplicable is the fact that the a is visible only on first stamper pressings but no longer visible on stamper II or IIIII pressings!

The third major type of error involved the duplicate use of, usually, short runs of catalog numbers. Several of these are shown below, as indicated in Kelly’s various catalogs. Such duplicate usage was quite frequent. In addition to the examples shown below, there are some instances of the same catalog numbers having been assigned to three different sets of recordings.

GRAMOFOON-ORKEST (ADRIAAN BLOKLAND) (The Hague)

90500 6909o - 7-07 La Bérolina, gezelschapsdans

90501 6910o - 7-07 Avec aplomb, marsch

90502 6911o - 7-07 Kreuz-Polka

90503 6912o - 7-07 Tendresse, caprice-mazurka

90504 6913o - 7-07 Pas de quatre

90505 6914o - 7-07 Bal-blanc, bostonwals

90506 6915o - 7-07 L'amour boiteux, polka-marsch (Fragson)

90507 6916o - 7-07 Marsch finale du Scala

90508 6917o - 7-07 Fringant, galop met bellen (Parès)

(The following numbers were duplicated by Hayes in 1918)

PICCADILLY ORCHESTRA (DAVID DE GROOT) (London)

90500' HO 4241ae 8- 7-18 In de Jordaan (Davids-Morris) B4506

90501' HO 4242ae 8- 7-18 Lettre à Armand (A Haagman) B4507

90502' HO 4244ae 8- 7-18 Radijs-wals (Davids-Morris) B4506

90503' HO 4246ae 8- 7-18 Viens au cabaret (A Haagman) B4507

90504' HO 4516ae 25-11-18 Ballgeflüster (Meyer-Helmund) B4508

90505' HO 4518ae 25-11-18 Si vous l'aviez compris (Denza) B4508

90506' HO 4521ae 25-11-18 Esclave d'amour (A Haagman) B4509

90507' HO 4523ae 25-11-18 Les gages d'amour (A Haagman) B4509

---

GUARDIA REPUBLICANA (Paris)

260249 6202h -08 La Brabançonne (Campenhaut) AE293

BANDA (K.u.K. INFANTERIE-REGIMENT No 51, FREIHERR VON DAVID (Vienna)

260250 15210u 29-9-09 Oesterreichische National Hymne AE293

BANDA (COLDSTREAM GUARDS) (London)

260251 y13277e 28-2-11 Rule Britannia (Arne arr. F Godfrey)(dir. ROGAN)

650093 AE294

260251 X Bb2170-1 2 0-11-22 do (dir. R G EVANS) AE294 (cancelled 1/4/24)

260252

BANDA MUNICIPAL DE BARCELONA (Barcelona)

260253 19044u 10-2-15 Canción del Rhin, Danza (Christiné) 650204 AE295

260254 19033u 5-2-15 Risuena, Mazurka (Oliva)

260255 19030u 5-2-15 El Ingles, Schottisch (Oliva) AE294

BANDA (ORCHESTER SEIDLER-WINKLER) (Munich)

260256 3926r -07 Heil dir im Siegerkranz, Marsch (Carey) 650086

BANDA (KRALEVE GARDE (ST BINICHOD)) (Belgrade)

260257 13660b -9-09 Srpska narodna himna (Fenka) 650205 AE295

BANDA (GUARDA REPUBLICANA) (Lisbon)

260258 1350ah 8-10-11 A Portugueza (AKeil)

BANDA (BLACK DIAMONDS) (London)

260259 11490e 18-3-10 Japanese National Hymn 650099

260260 10734e 17-9-09 Ahmed V, March (Turkish national hymn) 650094

BANDA (HARMONIE ORCHESTER) (Berlin)

260261 13253r 5-11-13 Deutschland über alles (Haydn)

The outbreak of the Great War evidently produced a disruption at this point, lasting seemingly for a year. The following catalogue numbers (260249 to 260261) were then duplicated by Barcelona, the reuse presumably being unintentional.

BANDA MUNICIPAL (Barcelona)

260249' 19034u 5-2-15 La chubona (Larruga) AE293

260250' 19037u 8-2-15 La patinadora, Polka (Suñé) AE293

260251' y13277e 28-2-11 not reused - see 260251

260252' 19045u 10-2-15 Tu has caido chagneton, Danza (Valverde)

260253' 19044u 10-2-15 not reused - see 260253

260254' 19033u 5-2-15 not reused - see 260254

260255' 19030u 5-2-15 not reused - see 260255

260256' 19038u 8-2-15 La discusion, Habanera (Furés)

260257' 19040u 8-2-15 Pepita, Polka (Oliva) AE295

260258' 19042u 10-2-15 El Xibarre, Vals-jota (Casademont) AE295

260259' 19028u 3-2-15 La fregoncilla, Polka (Furés) 650099 AE296

260260' 19029u 5-2-15 Impresionista, Schottisch (Casademont) 650094 AE296

260261' 19046u 10-2-15 Desemgañy, Tango (Casademont) 650133 AG24

---

LESOEUR (Amsterdam)

33265 5299a - 3-03 Le Coeur et la Main: Boléro (Lecocq)(mx 5299 DUTCH)

33266 5304a - 3-03 La Fille de Madame Angot: Coupletten (Lecocq)

(mx 5304 DUTCH)

33267 5298a - 3-03 Mam'zelle Carrabin (mx 5298 DUTCH)

33268 5302a - 3-03 La Poupée: Couplets du mannequin(Audran)

(mx 5302 DUTCH)

M A VASILEVA (St Petersburg)

33269 1274C - -03 Voulez-vous, Chansonette not used - see 23544

The following numbers were duplicated by Paris

Mlle CORTEZ

33265' 1542F - -03 Alceste: Divinités du Styx (Gluck)

33266' 1543F - -03 Pareil à la mer profonde (A Holmès)

33267' 1544F - -03 Mireille: Couplets (Gounod)

33268' [1545F] - -03 title not traced

33269' 1546F - -03 Orphée et Euridice: Plein de trouble (Gluck)

A notable error of this kind has been discussed by Alan Kelly relative to the multiple use of the catalog number 30095. Kelly lists 4 entries with this catalog number, as

MUSIQUE DE LA GARDE R�PUBLICAINE30090 3224- - 7-99 La Fille du Régiment: Fantaisie (Donizetti)30090 265F - -02 do 8000130090X 2532F - -03 do [reserve]30091 3226- - 7-99 Les flots du Danube, Valse (Ivanovici)30091' 264F - -02 do30092 3234- - 7-99 Schottish des pierrots (Lamotte)30092' 208F - -02 do 8009730092X 3203F - -04 do

The above listing is not quite correct, since a black label copy of G.C.-30092X bears the matrix number 208 clearly visible to the right of the spindle hole. Therefore the listing above should read

30092X 208F - -02 do 8009730092X’ 3203F - -04 do

The recording engineers did not always use all of the serial numbers allocated to them for any given recording size; most of them seemed to prefer to start a new recording series in a new location with a fresh group of matrix numbers. Thus, after Fred Gaisberg had ended his monumental series of recordings in the Sistine Chapel in the Vatican during the first week of April 1902, with serial number 1762, he began his next recording session in Rome with 1771, thus skipping eight numbers which were never used. A similar situation occurred when he had concluded the recording sessions in Milan with Enrico Caruso and others on April 11, 1902, with serial 1792; he followed this in Zürich three days later, where he used 1801 through 1809, and then continued in London with 1900. Such elisions were many and frequent during the acoustical era, particularly when a series of recording sessions in one city and country had concluded and the engineer moved on to another city or country. Of Fred’s entire suffix-b series from 100 to 20689, some 1,208 serial numbers were never used and at least another 1,222 matrices were destroyed, while an additional 115 were either rejected or used as test recordings. During his trip to the Caucasus in the summer of 1915, Fred says that he made some 480 recordings which were sent to Riga for processing but were never seen again! Thus some 3,000 serial numbers in the b suffix series were never used.

Additional suffixes seen on early matrix numbers, as described below, were most likely added by various plant technicians, possibly either to indicate the sequential flow of events in the manufacturing process, or to identify the various processing stations through which a particular black record or secondary stamper should pass. As early as March 1901 a much smaller Berliner logo had been introduced, and the Recording Angel trademark had been moved back to the left of the spindle hole, as seen in various figures below.

It should be noted here (as it probably has never been noted before or elsewhere) that many serial numbers found on early recordings of the Gramophone Company appear quite complex, and have been so designated by various discographers. Moreover, as Alan Kelly has observed, the serial number was probably entered onto the central area of the recording tablet following the actual recording, rather than before it. This is confirmed by the fact that most early matrix numbers were inscribed just inside the runoff area next to the grooved area. This would have been rather impossible if the matrix number had been entered before the recording had been made.

The earliest of such compound serial numbers seen by me is on a Berliner 2354 of February 4, 1899; at least three almost illegible suffixes follow the serial number 1173, as seen below

These additional suffixes appear to have been used solely by personnel at the Riga manufacturing plant, and are found only on discs recorded and issued in Russia, although the practice seems to have been cancelled following the introduction of the new letter/triplet system for matrix suffixes. They would have placed them there in order to designate the proper channels by which the recording tablets were to progress through the manufacturing process. Bennett lists a matrix number 195G-N-15 for G.C.-22492, recorded in Warsaw in June 1901. One sees such compound serial numbers as early as December 1901, on a Vialtzeva recording G.C.-23130 with the serial number 319X─F─2^z impressed plainly through the label. The compound serial number 440^z─Ao─2^z is seen on a February 1902 Battistini recording G.C.-52664, with identical additional suffixes on all other stamper II pressings from his first recording session in Warsaw. Some twenty recordings by the Russian baritone Polikarp D. Orlov, as well as fourteen by N.G. Svetlanov, and four by Dmitri Bukhtoyarov all show the additional suffix Ao-2^z, while groups by A M Labinsky, Leonid Sobinov and others show the characters –nB-15, accounting for nearly all of the 10-inch issues before the end of 1901. These expanded matrix numbers are seen in the runoff area as well as under the label, while the serial number and suffix can be seen through the label. They were used only on ten-inch recordings, and mainly between the time of their introduction in late 1901 and about February 1902.

They occur on the Chaliapin issues of February 1902, viz., 573x-N_3L-2_z on a first stamper pressing shown above, manufactured, interestingly enough, in the Hanover plant.

A first stamper pressing, probably processed after March 1902, of Nikolai Figner’s recording G.C.-22596x, shows the matrix number 369^z-T₂-2^z in the runoff area. Perhaps the most unusual and most unexpected group of additional suffixes can be seen under the label, flush within a raised ring, of a first stamper pressing recorded by Olimpia Boronat in St. Petersburg in January 1904 and processed at the Riga plant. These suffixes, L-2^z, can be seen inscribed well below the expected matrix number 1776 L. expected matrix number Note that the superscripts and subscripts are shown as seen on the discs themselves.

It is the author’s contention that the first number, perhaps combined with a possible superscript letter, was the actual matrix number entered during the recording session, while the additional notations were made by various employees and technicians at the Hanover plant. They are completely absent from the French Catalogue, and occur only rarely in the German Catalogue. The extremely complex 676z G 2z I I is found on a March 1903 recording G.C.-84018 made in Stockholm, probably by Franz Hampe, the final z having been used by him until late 1903 or early 1904. [Author’s note: this may be confused with a 1901 recording made in Paris by Fred Gaisberg, including his G.] To quote John Ward (see Bibliography), who corroborates this opinion, "they are irritating additions which can be ignored."

It is unfortunate that John Bennett confined his listings to vocal recordings only. It is highly likely that compound matrix numbers such as those described above would be found on many other categories of recordings made and issued between the introduction of 10-inch discs in April 1901 and the end of February 1902.

Catalog Numbers

When the finished recording tablets arrived at the processing plant, catalog numbers were assigned according to the criteria outlined above, albeit frequently in random order. The table below shows the Weekly Returns from Italy for Friday, April 11, 1902 and for the first two days of the week ending on December 7, 1902. The matrix numbers are in sequence, while the catalog numbers are out of sequence.

Recording Date	Matrix Number	Catalog Number	Artist		Recording Date	Matrix Number	Catalog Number	Artist
April 02	1771	52356	Oxilia		30 Nov 1902	2861	52410	De Lucia
April 02	1772	not issued	Oxilia		30 Nov 1902	2862	52435	De Lucia
April 02	1773	not issued	Oxilia		30 Nov 1902	2863	52411	De Lucia
April 02	1774	52357	Oxilia		30 Nov 1902	2864	52427	De Lucia
11 April 02	1775	53259	Pinto		30 Nov 1902	2865	52436	De Lucia
11 April 02	1776	53233	Pinto		30 Nov 1902	2866	52412	De Lucia
11 April 02	1777	53238	Pinto		30 Nov 1902	2867	52413	De Lucia
11 April 02	1778	not issued	?		30 Nov 1902	2868	52414	De Lucia
11 April 02	1779 R	52358	Oxilia		30 Nov 1902	2869	52415	De Lucia
11 April 02	1780	53234	Pinto		30 Nov 1902	2870	52416	De Lucia
11 April 02	1781	53240	Pinto		30 Nov 1902	2871	52438X	Caruso
11 April 02	1782-BG	53232X	Pinto		30 Nov 1902	2872	52439	Caruso
11 April 02	1782-nB	52378	Caruso		30 Nov 1902	2873	52369X	Caruso
11 April 02	1783	52344	Caruso		30 Nov 1902	2874	52417	Caruso
11 April 02	1784	52369	Caruso		30 Nov 1902	2875	52440	Caruso
11 April 02	1785	52345	Caruso		30 Nov 1902	2876	52418	Caruso
11 April 02	1786	52346	Caruso		30 Nov 1902	2877	52441	Caruso
11 April 02	1787	52347	Caruso		1 Dec 1902	2878	53259	Bresonnier
11 April 02	1788	52379	Caruso		1 Dec 1902	2879	52443	Caruso
11 April 02	1789	52348	Caruso		1 Dec 1902	2880	52419	Caruso
11 April 02	1790	52349	Caruso		1 Dec 1902	2881	52420	De Lucia
11 April 02	1791	52368	Caruso		1 Dec 1902	2882	52442	Caruso

The following table shows the same information arranged according to the assigned catalog numbers. The lower portions of the table show the numbers missing from the Weekly Returns listed above. These were assigned to recordings made in either March 1902 or November-December 1902. These indicate clearly that catalog numbers were assigned quite independently from the matrix numbers of the wax tablets received.

Recording Date	Matrix Number	Catalog Number	Artist		Recording Date	Matrix Number	Catalog Number	Artist
April 02	1772	not issued	Oxilia		30 Nov 1902	2871-R	52348X	Caruso
April 02	1773	not issued	Oxilia		30 Nov 1902	2873-R	52369X	Caruso
11 April 02	1778	not issued	?		30 Nov 1902	2861	52410	De Lucia
11 April 02	1783	52344	Caruso		30 Nov 1902	2863	52411	De Lucia
11 April 02	1785	52345	Caruso		30 Nov 1902	2866	52412	De Lucia
11 April 02	1786	52346	Caruso		30 Nov 1902	2867	52413	De Lucia
11 April 02	1787	52347	Caruso		30 Nov 1902	2868	52414	De Lucia
11 April 02	1789	52348	Caruso		30 Nov 1902	2869	52415	De Lucia
11 April 02	1790	52349	Caruso		30 Nov 1902	2870	52416	De Lucia
April 02	1771	52356	Oxilia		30 Nov 1902	2874-R	52417	Caruso
April 02	1774	52357	Oxilia		30 Nov 1902	2876-2/W	52418	Caruso
11 April 02	1779 R	52358	Oxilia		1 Dec 1902	2880	52419	Caruso
11 April 02	1791	52368	Caruso		1 Dec 1902	2881	52420	De Luca
11 April 02	1784	52369	Caruso		30 Nov 1902	2864	52427	De Luca
11 April 02	1788	52370	Caruso		30 Nov 1902	2862	52435	De Luca
11 April 02	1782-nB	52378	Caruso		30 Nov 1902	2865	52436	De Luca
11 April 02	1782-BG	53232X	Pinto		30 Nov 1902	2872-2/W	52439	Caruso
11 April 02	1776	53233	Pinto		30 Nov 1902	2875-2/W	52440	Caruso
11 April 02	1780	53234	Pinto		30 Nov 1902	2877-2/W	52441	Caruso
11 April 02	1777	53238	Pinto		1 Dec 1902	2882-R	52442	Caruso
11 April 02	1781	53240	Pinto		1 Dec 1902	2879-2/W	52443	Caruso
11 April 02	1775	53239	Pinto		1 Dec 1902	2878	53259	Bresonnier
Mar 02	1740	52350	Malesci		1 Dec 1902	2881	52421	De Luca
Mar 02	1741	52351	Malesci		2 Dec 1902	2883	52422	De Luca
Mar 02	1742	52352	Malesci		2 Dec 1902	2884	52423	De Luca
Mar 02	1745	52353	Malesci		2 Dec 1902	2885	52424	De Luca
Mar 02	1746	52354	Malesci		2 Dec 1902	2886	52425	De Luca
Mar 02	1743	52355	Malesci		2 Dec 1902	2887	52426	De Luca
Mar 02	1684	52359	Caffetto		Nov 02	2841	52428	Garbin
Mar 02	1634	52360	Oxilia		Nov 02	2838	52429	Garbin
Mar 02	1633	52361	Oxilia		Nov 02	2844	52430	Garbin
Mar 02	1644	52362	Caffetto		Nov 02	2842	52431	Garbin
1902	869x	52363	Felix		Nov 02	2839	52432	Garbin
Mar 02	1683	52364	Caffetto		Nov 02	2840	52433	Garbin
Mar 02	1685	52365	Caffetto		Nov 02	2843	52434	Garbin
Mar 02	1687	52366	Girardi
Mar 02	1713	52367	Gravina		3 Dec 1902	2899	52437	De Lucia
					3 Dec 1902	2900	52438	De Lucia
Mar 02	1705b	52371	Sammarco
Mar 02	1706b	52372	Sammarco
Mar 02	1707b	52373	Sammarco
Mar 02	1709b	52374	Sammarco
Mar 02	1708b	52375	Sammarco
Mar 02	1686b	52376	Caffetto
1902	1011x	52377	Schrodter

The table below shows the above catalog numbers in sequence, to indicate the randomness of the recording dates and the assignment of catalog numbers.

Recording Date	Matrix Number	Catalog Number	Artist		Recording Date	Matrix Number	Catalog Number	Artist
11 April 02	1783	52344	Caruso		30 Nov 1902	2868	52414	De Lucia
11 April 02	1785	52345	Caruso		30 Nov 1902	2869	52415	De Lucia
11 April 02	1786	52346	Caruso		30 Nov 1902	2870	52416	De Lucia
11 April 02	1787	52347	Caruso		30 Nov 1902	2874-R	52417	Caruso
11 April 02	1789	52348	Caruso		30 Nov 1902	2876-2/W	52418	Caruso
30 Nov 1902	2871-R	52348X	Caruso		1 Dec 1902	2880	52419	Caruso
11 April 02	1790	52349	Caruso		1 Dec 1902	2881	52420	De Luca
Mar 02	1740	52350	Malesci		1 Dec 1902	2881	52421	De Luca
Mar 02	1741	52351	Malesci		2 Dec 1902	2883	52422	De Luca
Mar 02	1742	52352	Malesci		2 Dec 1902	2884	52423	De Luca
Mar 02	1745	52353	Malesci		2 Dec 1902	2885	52424	De Luca
Mar 02	1746	52354	Malesci		2 Dec 1902	2886	52425	De Luca
Mar 02	1743	52355	Malesci		2 Dec 1902	2887	52426	De Luca
April 02	1771	52356	Oxilia		30 Nov 1902	2864	52427	De Luca
April 02	1774	52357	Oxilia		Nov 02	2841	52428	Garbin
11 April 02	1779	52358	Oxilia		Nov 02	2838	52429	Garbin
Mar 02	1684	52359	Caffetto		Nov 02	2844	52430	Garbin
Mar 02	1634	52360	Oxilia		Nov 02	2842	52431	Garbin
Mar 02	1633	52361	Oxilia		Nov 02	2839	52432	Garbin
Mar 02	1644	52362	Caffetto		Nov 02	2840	52433	Garbin
1902	869x	52363	Felix		Nov 02	2843	52434	Garbin
Mar 02	1683	52364	Caffetto		30 Nov 1902	2862	52435	De Luca
Mar 02	1685	52365	Caffetto		30 Nov 1902	2865	52436	De Luca
Mar 02	1687	52366	Girardi		3 Dec 1902	2899	52437	De Lucia
Mar 02	1713	52367	Gravina		3 Dec 1902	2900	52438	De Lucia
11 April 02	1791	52368	Caruso		30 Nov 1902	2872-2/W	52439	Caruso
11 April 02	1784	52369	Caruso		30 Nov 1902	2875-2/W	52440	Caruso
30 Nov 1902	2873-R	52369X	Caruso		30 Nov 1902	2877-2/W	52441	Caruso
11 April 02	1788	52370	Caruso		1 Dec 1902	2882-R	52442	Caruso
Mar 02	1705b	52371	Sammarco		1 Dec 1902	2879-2/W	52443	Caruso
Mar 02	1706b	52372	Sammarco
Mar 02	1707b	52373	Sammarco		11 April 02	1782-BG	53232X	Pinto
Mar 02	1709b	52374	Sammarco		11 April 02	1776	53233	Pinto
Mar 02	1708b	52375	Sammarco		11 April 02	1780	53234	Pinto
Mar 02	1686b	52376	Caffetto		11 April 02	1777	53238	Pinto
1902	1011x	52377	Schrodter		11 April 02	1775	53239	Pinto
11 April 02	1782-nB	52378	Caruso		11 April 02	1781	53240	Pinto
30 Nov 1902	2861	52410	De Lucia		1 Dec 1902	2878	53259	Bresonnier
30 Nov 1902	2863	52411	De Lucia		April 02	1772	not issued	Oxilia
30 Nov 1902	2866	52412	De Lucia		April 02	1773	not issued	Oxilia
30 Nov 1902	2867	52413	De Lucia		11 April 02	1778	not issued	?

When it was necessary for an artist repeat a recording of a selection for one reason or another, usually but not always in the same recording session, the earliest custom was to assign the next subsequent matrix numbers to any additional recordings or "takes," and to note these in the Weekly Returns and ledgers. When one take was selected in preference to another for processing and issue, a catalog number was then assigned. If for any reason an additional "take" was to be issued, it was assigned the same catalog number, but a suffix was added to indicate which "take" had been used. Thus second, third, fourth, and fifth takes were assigned a catalog number with the suffix X, Z, W, and Y, respectively. In addition new serial numbers were usually assigned to additional takes. Note that the letter following the catalog number on the surface of the disc itself was a capital, while that printed on the label was in lower case. (Not an easy job for a typesetter in those days!) Multiple uses of letters have also been observed.

It may be useful at this point to review the major difference between matrix or serial numbers and catalog or record numbers. The first issued discs of the Gramophone Company were identified by a catalog number, to which was attached a prefix as indicated above. In the listings of the first 406 recordings in the "original" series, some twenty recordings are listed with catalog numbers having the suffix X, in one instance with Y. This practice has been seen as late as December 1913, although the new matrix system, in which the matrix number rather than the catalog number was modified to indicate additional takes, had been introduced in early 1904. From as early as March 1901 a second take had been indicated by adding the suffix ½ to the original matrix number. A third take was indicated by the suffix ¾. An alternate system used ½, ⅓, and ¼ as suffixes, as shown below. Alan Kelly has pointed out that the presence of such suffixes may not always indicate additional takes. The fraction system was used as late as June 1919 on 10-inch recordings. Needless to say, numerous erroneous assignments of these suffixes occurred, some matrix numbers even being given a fractional suffix and then assigned to a different artist and selection altogether! The interested reader should refer to the Perkins, Kelly, and Ward article in the Bibliography for other variant usage.

The matrix numbers shown above are from the same recording, the original number being that on the left used by the Victor Talking Machine Company from the original Gramophone Company shell. The number on the right was reentered onto a secondary stamper III used for a Gramophone Company pressing.

In addition to the catalog number, each issued record was supposed to have a serial number, which was restarted from 1 each day and when taken with the recording date constituted the matrix number. This practice was changed to a continuous series when the "unlettered" series began on November 1, 1898, probably due to the occurrence of these second and third takes. Thus the catalog number indicates the where or location, region or language, i.e., the what or category or type of the selections and often the who, i.e., the number and gender of the artists, while the serial or matrix number indicates the when and how of a recording. The label itself of course tells the rest of the story, i.e., the location, the selection(s), the artist(s), and usually the vocal range or instrument.

Fortunately, alternate takes are recognized more commonly and easily from the new serial number assigned to the selection, since the "take" suffix for the catalog number does not always appear on the label. A notable case in point involves the two second takes made by Enrico Caruso in his second session on November 30, 1902, for which he had agreed to re-record two arias from the first session that were considered to be unsatisfactory. These were Celeste Aida, with the original matrix number 1784, and Dai campi, dai prati, with the original matrix number 1789. [Author’s note: Freestone and Drummond indicate that the first Celeste Aida was a far better rendition than the retake. However, the first take survives in pressings from two known stampers, indicating that between 700 and 1,000 discs were probably issued, as does the first recording of the Mefistofele aria.] The second takes were assigned serial numbers 2871-R and 2873-R, respectively, both being recorded by Belford (frequently misquoted as Bedford by many writers) Royal, who assisted Will Gaisberg in this second recording session.

The three records above each show the correct catalog number on the record surface at the 12 o’clock position, including the X suffix. However, the Dai campi, dai prati retake was first issued with the label on the left which had been used for the first take, showing the original selection title and the catalog number 52348. Secondary stamper pressings were issued with a new label showing the composer Boito’s name and the catalog number 52348x, while the Celeste Aida retake still has the original catalog number 52369 on the issued label. Nevertheless, some collectors and dealers still list first, second, or third stamper pressings of the Aida retake as pressings of stampers X I, X II, or X III of the first take, while ignoring the earlier matrix number. Although the labels used for the latter retake appear to have been remainders from the previous printing for the first take, three different label designs are known for this issue, up to stamper IIII pressings, indicating that the printers were either unaware or uninformed of the fact that these discs were pressed from retakes, and that the catalog number on the label should have been 52369x.

Following the introduction of the so-called "five-stage" process attributed to Eldridge Johnson, it became necessary to re-enter both the catalog number and the matrix number in the runoff area, usually at 12 o’clock and 6 o’clock, respectively. Occasionally this led to errors resulting from the entry of a new form of the catalog number or matrix number over the one used for the previous stamper, as seen in the two images below. It is apparent that the original catalog number on the right is identical with that on the left.

 

Note the change in font and spacing from stamper IIII to stamper VI above, and in those below from stamper VII to stamper IX.

Note that the two catalog number entries below are not identical, as the 2 on the right is slightly higher than that on the left

 

The catalog numbers below appear to be identical with that from stamper III above. Note the double entries on the right.

 

 

On the other hand, one can see from the figure below that the identical embossed catalog number was used on at least three different stampers for Caruso’s first G&T recording.

 

 One unusual assignment of catalog numbers is seen in regard to the Sarasate recordings made in Paris in 1904 and issued originally with catalog numbers in the French series. Six of the ten recordings were later issued with German catalog numbers, as shown below in a modified extract from Kelly’s French and German catalogs. It is generally believed that such catalog numbers were assigned when recordings were reissued for an audience whose language differed from that of the country of origin. The double-sided reissues in the E series are shown, as well as known Victor issues.

French	PABLO DE SARASATE (p) (Paris)			German
37929	4262°	-7-04	Caprice basque Op 24 (Sarasate) Vic 63168-B
37930	4263o	-7-04	Zigeunerweisen Op 20 (Sarasate), pt 1 Vic 63167-A E329 EW2	47962
37931	4258°	-7-04	Partita No 3 BWV1006: Prélude (Bach), unaccompanied
			Vic 67903 E183 EW3
37932	4259°	-7-04	Caprice jota Op 41 (Sarasate)
			Vic 67900 ER76	47966
37933	4260°	-7-04	Tarentelle Op 43 (Sarasate)	47965
			Vic 62111-A 67904 E183 EW3
37934	4261°	-7-04	Miramar "Zortzico" Op 42 (Sarasate)
			Vic 62110-B Vic 52708 ER75	47964
37935	4264o	-7-04	Zigeunerweisen Op 20 (Sarasate), pt 2
			Vic 63167-B 67902 E329 EW2	47963
37936	4265°	-7-04	Habanera Op 21 No 1 (Sarasate) Vic 62110-A 67905 ER76	47967
37937	4266o	-7-04	Zapateado, Danse des souliers Op 23 No 2 (Sarasate)
37938	4267°	-7-04	Nocturne Op 9 No 2 (Chopin-Sarasate)

 

     

French Sarasate Issues

The original labels are shown above. The pair of labels below from a double-sided disc show German catalog numbers, in spite of their having been recorded in Paris some five years before these records were issued. Note the embossed Angel trademarks, and the language/region indicated as German.

     

German Sarasate Issues

Alan Kelly has uncovered the following letter written by Will Gaisberg in 1910, as follows:

 

Laboratory, December 13th, 1910.

Gentlemen,

I want to make a slight change in the serials of the experts’ numbers in London, to show which expert who happens to be recording in London, has made the record.

At the present time we practically use the one serial in London, viz., my own serial. In the future, I wish to add, besides the letter "e" to the London ten inch, or the letter "f" to the London twelve inch, the letter or letters of the expert in charge. For example, if Mr Arthur Clarke is making the records in London, the ten inch records would read thus

y 46926e

In this way the London numbers will always remain the same, the only difference being that the expert in charge adds his letter to the present serial now in use, so that if he happens to go away, and another expert takes his place, there will be no variation in the numbers, except the letter which goes in front will be changed to that of the expert who makes the records.

Will you let me know if this will in any way make trouble to the factory. If not, it will be a great convenience to the office here.

Very truly,

W.C.G.

This letter was in response to the practice at that time for all engineers to use the suffixes e and f for all London recordings. The modified system proposed by Will allowed one to identify the actual recording engineer, who by that date numbered fourteen or more. The prefixes were the same as the suffixes assigned as new engineers were employed. Thus Hancox had aa, ab, and ac, Edmund Pearse was given ad, ae, and af, while Hugh Murtagh, A. S. Clarke, and Fred Gaisberg shared ag, ah, and ai/aj. George Dillnutt used ak, al, and am, Beckwith was given am and ar, and John Smoot used sm and sn.

Previous page

Next page

Return to Book Index