In many of the items in our files, the american use of the hash sign or sharp (#) has been adopted (rather than the italian "No.") to denote numbers, especially serial numbers: Also, in order to make long numbers easier to read, several readers have asked me to insert commas after the thousands, although these do not normally appear on the pianos. Hallmarks for silver and porcelain, etc. are published widely in many books, but the same kind of information is not available for pianos, and if there are no datemarks, serial numbers are one option, but dating pianos by their serial numbers is notoriously unreliable, partly because so much misleading information has been published. Datemarks are a much more accurate and reliable guide to the age of an instrument.

Dating a piano by its number depends on four things:

1. The piano must have a number! Sometimes, it is so well hidden that you may not find it. (If you do find just one, don't assume that it is necessarily the right one.)

2. The numbers must be published. Many are simply not available, although my files include many dates of numbers which are not published anywhere else. Books which give number dates include Michel's Piano Atlas, (which later became the Pierce Piano Atlas) The Europe Piano Atlas, and the Musicians' Piano Atlas and its supplement, for which I supplied appendices with action & key numbers etc..

I am grateful to the Castle Museum, York, for copies of a list of piano serial number dates compiled by Mr D.R. Homan, a piano tuner from Scunthorpe. Most of these lists end around 1935-6, so we imagine that they were collected around that time. Some agree with other published numbers, some do not, which brings me to...

3. The published numbers must be CORRECT! A few years before I became involved with pianos, N.E. Michel published the first "Piano Atlas" of serial numbers: Some of the numbers he published are wrong or misleading: Early numbers for Bluthner, Bord, and Brinsmead, do not correspond with the instruments. Neither do Bristol pianos, which he wrongly attributes to Duck, Son & Pinker. Clementi and Collard dates are a complete nonsense. The Erard entries make no allowance for the fact that many instruments from London bear a separate sequence of numbers. Monington's numbers only show one of the two sets. Wornum's numbers are misleading because they do not give the Piccolo Pianoforte numbers. These and other mistakes have been perpetuated by other books reprinting them without evidence to support the information, but then it is a difficult trap to avoid, and everyone quotes published information sometimes, taking it on trust. The Musicians' Piano Atlas published revised entries for Jarrett & Goudge, based on nothing more than my vague suggestion that Michel's numbers were about forty years out!

4. There are often different numbers on the casework, keys, wrestplank, soundboard, iron frame, and action: Which is the correct number?

NUMBERS: Which is which?

Some people solve this problem simply by choosing the number which best suits their purposes, or makes the piano oldest! Often, the serial number may be imprinted into the wood of the casework, and sometimes it may be conveniently located on the top edge of the end of an upright piano, but this is by no means reliable, and grand case numbers are often harder to find, being hidden or underneath: Remember too the dangers of crawling underneath a grand - check that the legs are safe first.

By the late 1800s, removable case parts were often imprinted with the last 3 digits of the case number, so that the parts could be kept together during manufacture, and this helps to confirm which is the case number. Steinways marked even the tiniest pieces of moulding in this way. Coincidences can happen, and a 3-figure model number may happen to coincide with the last 3 digits of the case number, as it did on an old Collard square piano I owned. Pleyel pianos of the 1840s would have the case number on the case, action, keyblocks, and various other places, and their published numbers seem about right.

Having said all that, case numbers are not necessarily the published serial numbers, which may be painted onto the iron frame, or imprinted on the soundboard. Some of these belly numbers (such as those of Bechstein) run in separate sequences for uprights and grands, and exist in addition to the main series. Some makers, such as Wornum and Broadwood, had separate numbers for different models, as well as a main series, and odd things can happen.

Wornum's Piccolo Pianofortes have separate numbers, but as more and more of his production was devoted to these, their numbers almost began to catch up with the main number sequence.

Another point worth remembering is that because some conventions in numbering developed gradually over a period of many years, they are less applicable to early instruments.

A problem which occurs with Collards is that although the case number is sometimes confirmed by being repeated in several places, it is not the published number!

NUMBER OR DATE?

Numbers for most makers will go through a period when they are in four figures, and may resemble a year, so it is important to realise that pianos are hardly ever marked with just a year.

Some Misleading Numbers:

1792 John Broadwood square piano in the Mobbs Collection, Bristol, has the number 1931 - definitely NOT a date!

Circa 1805 Thomas Tomkison, London, square piano #1965 in the Smithsonian.

1833 John Broadwood & Son upright piano #1778 in the Mobbs Collection, Bristol.

1931? Dagmar piano #85,425, keys marked 1931, James Smith & Son, sole agents, 76/72 Lord Street Liverpool. Also marked JS&S September 1947, so perhaps 1931 is a just number?

Circa 1931 Two G.Ajello & Sons pianos around #28,500 are dated quite reliably by Malcolm action numbers, in spite of the mark 2.10.9 on the iron frames.

1931 Monington & Weston Piano #50,347 has 2.12.9 on the frame, similar to the example above, so perhaps it is a date code for 1931. Number on bracing #15,502.

NUMBERING PECULIARITIES

It might seem logical that serial numbers should, by their very definition, run in a single, unbroken series starting from the first piano a company makes, and therefore represent the number of pianos that have been made up to any given point in the list. Steinweg, for example, made 482 pianos before he emigrated to New York, and these were numbered from 1 up to 482, so he started at 483 in New York. However, although several authors have estimated production figures for various firms on the basis of serial numbers, it is not always that simple:

Around 1910, Murdochs' ad boasted "fifty thousand pianos in use" and perhaps Spencers, whom they took over, had made that many, but it is quite likely that the statement was based on their serial numbers. Some makers would (like Steinweg) started from a good, honest #1, and numbered every piano as they went, but some preferred to start at a more generous figure.

NUMBERS - THOUSAND PER YEAR

Certainly, quite a number of firms seem to have opted for allotting a thousand numbers per year, even if their production figures were nowhere near that many. Indeed, perhaps it was based on the assumption that they would never need any more than that. This convention can be seen in many lists of serial numbers, but since most of these were originally published in Michel's Piano Atlas, I thought for a while that they might be Michel's own estimates, until other examples began to appear in my files from quite independent sources, such as the pianos themselves.

Clementi's early pianos seem to work out at about a thousand numbers per year from their factory fire, a sequence which was continued for a while when Collard & Collard took over the firm. In 1883, Bluthners, who had been allowing a thousand per year, increased this to fifteen hundred, to cope with their huge and growing output. In 1934, Eavestaffs' first back-action minipiano was made. The serial numbers, again, seem to be based on a thousand per year.

An interesting angle here is to see how few pianos whose numbers go up in thousands have a high number of hundreds on the end. Strip away the thousands, and you'll find relatively few which have a remainder over five hundred.

Strangely, a Piano Times article on Chappells says their piano production had risen to one hundred per week (5,200 per year) by 1922, yet in that period, the serial numbers still only went up by a thousand per year! Copies of warranty labels show that Chappells' "Elysian Piano" numbers also went up a thousand per year, and used date codes within the serial numbers, by placing the last 2 digits of the year at the beginning of the number, so 24,000 meant 1924. Ajello (Manchester) on the other hand, placed the two digits of the year at the end, which has a peculiar effect on the sequence of numbers: If I rounded these down, they would give no clue to the date. By 1928, Marshall & Rose serial numbers were counted in steps of a thousand per year, with the first 2 digits also representing the year, so #28,000 would be 1928, like the Chappell Elysian numbers.

NUMBERS OF ANCILLARY TRADES

Keys and actions are usually made by separate firms, and have their own numbers, which can sometimes be a guide to the date of a piano. I can often help with dates of these too. It is not usually possibly to trace the history of an individual piano, but I often have more information than the makers themselves, and I may be able to provide a useful cross- reference by looking up action & key numbers, frame markings, or historical data such as name changes, dates of addresses, etc..

NUMBERS OF MODELS

Model numbers rarely exceed 3 digits, whereas serial numbers are usually much bigger. Many makers use model numbers, and these may tell us several things about the piano: German firms like Bechstein and Grotrian used the size of the piano in centimetres - the length of a grand, or the height of an upright. Bentleys used the number of notes, so a Concord 88 had 88 notes. Berrys did a similar thing with the number of octaves, and followed it with two digits of the year it was introduced, so the Berry 758 was a 7-octave model introduced in 1958, although some models continued to made for years. Some early Collard & Collard pianos have a 3-figure number which may represent the model. Spencer-Murdoch pianos have what seems to be a model number at right-angles to the serial number, in smaller figures.

NUMBERS OF DEALERS' STOCK

Stock numbers often add to the general confusion, and unless they are preceded by the dealer's initials, it is difficult to know what they are. Sometimes, they are preceded by a letter or letters representing the branch, such as "MP" for Manor Park, or (more commonly) "HO" for Head Office. The letter P (for piano) may also appear in the case of a firm which sold other goods as well.

I have some individual dates of stock numbers on file, as well as a few lists, and although these are incomplete, they are often useful.

Circa 1910 Adlon upright #15,100 supplied by W.H. Barnes, Oxford Street, London W.1. also has 2391 & P16363 on the case, and the main number is repeated on the back bracing, inside the bottom of the piano.

NUMBERS OF ACTIONS

The following list gives approximate numbers of various action makers in thousands. This information has been updated since some of it was included in my appendix to Musicians Piano Atlas.

For further information, see the separate entries in our files.

Schwander

1885     225

1886     240

1887     255

1888     270

1889     285            Langer

1890     300              63 

1891     315              71 

1892     330              79 

1893     345              87 

1894     360   Kohler     95 

1895     375     50      103 

1896     387     60      111 

1897     400     70      119 

1898     415     80      127     Schutze & Freund

1899     430     90      136     430

1900     445    100      144     445

1901     460    120      153     460

1902     475    140      162     475

1903     490    160      173     490

1904     510    180      185     510

1905     530    200      210     530

1906     550    240      245     550

1907     570    280      280     570

1908     595    322      295

1909     620    348      310

1910     645    375      335

1911     670    424      360

1912     690    473      385

1913     700    488      410

1914     710    503      425

1915     727    520      440 

1917            530      470     Morgenstern & Kotrade 

1918            540      480     181 

1919            550      490 

1920            560      500 

1921            581      530             Malcolm

1922            586      570  Lexow      24

1923            590      615   616       30

1924            595      643   625       37

1925            600      665   641       63

1926            615      690   657       73

1927            715      672             83

1928            730       97

1929            755      102

1930            780      108

1935            121

1940            125

1945            129

The numbers for Schutze & Freund were arrived at quite independently from the Schwander numbers, but they are identical.

I would be pleased to hear from anyone who has dated any piano numbers by using this list.