By Roger Beardsley


As the opening paragraph of the article implies, the frequently repeated, but uninformed theories as to how the duplication of metal parts was achieved, were the driving force behind its writing. However, its original context meant that only stampers made from wax masters were being considered, and before metal to metal plating was successfully carried out using a separating film.

Before wax, the acid-etched zinc master was used. Metal negatives were obviously made from those metal masters, so the question arises as to why, if that were possible, could they not make duplicates from the metal shells taken from the wax.

The answer lies in the process and some basic metallurgy. After the zinc master was plated with copper in an alkaline bath, the zinc would be removed by sulphuric acid, which does not attack the copper, leaving the copper negative ready for pressing records. Of course, this was a one-trip process, and is the reason why when a recording was deemed likely to be popular beyond the pressing capacity of a single stamper, multiple takes were made of the same selection.

It also appears that a process involving lead and graphite as a separating film was tried. This was already in use in the photographic industry, but full details are being researched.

Putting the record straight about the replication of metal parts before metal-to-metal plating was introduced

This article was prompted by the many enquiries as to the differing and erroneous versions and explanations written of the process that the Gramophone Co. used in making multiple stampers.

We need first to examine the recording process from the beginning, albeit briefly.

The making of lateral-cut 78s began with a wax blank (ignoring the acid-etched Berliners of course). The sound made by the performer was converted either mechanically in the case of recordings made before 1925, or electro-mechanically for those after 1925, into vibrations that caused a cutting stylus to vibrate in sympathy, and in a lateral plane in a line to the axis of a rotating wax blank. When that stylus traversed the wax, a groove was cut that moved from side to side, mimicking the original vibrations - it is thus a modulated groove. When played back, the groove modulations are converted either mechanically or electrically into sound.

The original wax master that resulted from the recording process was far too soft to allow replay without severe damage to the delicate grooves, and in any case the recording needed to be reproduced in quantity.

The first stage in the replication process was to make the wax electrically conductive. This was achieved by dusting it with the finest available graphite. It was then electroplated, and when the resulting copper deposition was peeled from the wax, a negative image of the wax was obtained which had ridges, instead of grooves. This was called a 'first shell'. Using this, it was possible to stamp out records in a press, but it would not last for ever, and when worn out, perhaps after 1000 pressings (1), no more records could be made. Frequently it was possible to re-plate the wax to produce a 'second shell', but it was not always successful, and of course for a very popular record, there was a need for mass replication of shells for use as stampers.

The answer to this problem, was solved by a process for making duplicate shells. The original first shell was placed in what was called a dubbing pan (the origin of the word 'dubbing' to indicate copying). A special mixture of various waxes and ultra-fine plaster of Paris was then poured over the shell. When set, it resulted in a duplicate (positive) of the original wax. This could then be plated to produce a (negative) stamper for making records, but left the first shell unworn and (usually) undamaged. The process could be repeated whenever a fresh stamper was required. In theory, providing the first shell was not damaged by incorrect handling, the first stamper originating from this process would be just as good as the 50th stamper. It all depended upon how well the dubbing process was carried out.

The system was not perfect, tending to result in greater surface noise, and a lack of definition if not done carefully (there are other causes for those ills, but that is not relevant here).

From quite early on, there was a realisation that copy (duplicated/dubbed) shells were not generally as good as a shell made from the wax (3). See below.

There is much fascinating information in the EMI Archive concerning the process, with details as to the exact composition of the waxes used.

From a least the end of 1904, the letters "CO" were generally used to indicate a dubbed or duplicated shell (2).

Note, the later meaning of the word 'dubbing' as a transfer is not applicable here.

In a letter dated 25th November 1904, there is an interesting reference to making 'special' duplicates of 'big Artistes' such as Melba. In making these, it appears special care was taken by the engineer Royal, and that he had sometimes had at least 12 attempts at making duplicates before getting a good one.

In a letter dated June 1903 (and referring to Tamagno's records), it is stated clearly that the first shell is to be kept for making duplicates and the second is to be used for pressing. It is also apparent from this letter, that at least five shells could sometimes be obtained from the original wax. It is also mentioned that the original of 3001b (Esultate!) became defective when taking the third shell. It is made quite clear that they are not referring to duplicated (dubbed) shells.

It has been stated in various places that copy metals were made by plating specially prepared pressings. This is utter nonsense and has no basis in fact anywhere in the EMI Archives. However the dubbing process outlined here is well documented. Only the pirates plated pressings, and the results were dreadful. IF the proponents of this can produce any evidence, apart from hearsay, I for one will be most interested to see it.

Of course, eventually the metal-to-metal plating system was perfected, and was certainly in widespread use by 1918, although not everywhere (4).

© Roger Beardsley, Historic Masters Ltd.


  • 1. Letter dated 25th November 1904 from the English Branch of G & T to Head Office. 1000 pressings is quoted as probable life of a shell.
  • 2. Letter dated 3rd January 1905 from G & T to the Milan office. Reference to those records marked with "CO".
  • 3. Letter dated 25th November 1904 from the English Branch of G & T to Head Office. Reference to using the second shell to press records and to only press from a duplicate of the first shell, if the second shell breaks down.
  • 4. Letter to Milan Branch 10th August 1918.


  • Wax - the original recording made on a wax tablet
  • Shell - metal negative of the wax obtained by electroplating
  • First shell - the first metal negative obtained by electroplating
  • Second (and subsequent) shell - negative obtained by electroplating the wax for a second (or more) time if it survived the first plating.
  • Duplicate shell - a metal part obtained via the dubbing process.
  • Dubbing - a method of copying the original (or a subsequent) shell. N.B. this is not the same as dubbing by transferring a pressing to produce a new master.
  • Stamper - a shell or other negative, but prepared for use in a press.