Mechanical Music Registry Project
National Automatic Music Company / Automatic Musical Instrument Company
This National Automatic Piano rollography project began circa 1984, which by October of 1985 had progressed to the place whereby the late Ed Sprankle produced an 11 page rollography report that listed the following contributors: Willard E. Burkhardt, Claude Garibotti, Mike Montgomery, John Perschbacher, and Charlie Smallwood. As of this writing (September, 2018), Willard E. Burkhardt, Mike Montgomery, Charlie Smallwood, and Ed Sprankle are known to be deceased. The whereabouts of Claude Garibotti are unknown, other than he at one time had a sizable collection of National music rolls. John Perschbacher continues to be active in his happy pursuit of mechanical music artifacts, music rolls, and coin pianos.
These men worked together on the project as friends, with no one officially in charge, but Ed Sprankle took the lead in making up the actual rollography report. There were many delays in getting the report completed. Nevertheless, through sheer persistence the completed rollography report finally emerged. The keen foresight by these key people is honored herewith by recognizing them for their effort, and for being the foundational basis of this National Automatic Piano rollography, with this author contributing another little bit by providing a modern database platform to further the efforts of the founding fathers. Here is Ed Sprankle's introductory letter for the first National rollography report:
October 20, 1985
TO: Bill Burkhardt
Claude Garibotti
Mike Montgomery
John Perschbacher
Charlie Smallwood
FROM: Ed Sprankle, Oakland, California
Here’s the NATIONAL info that has accumulated so far. Some of the delay in getting this done was due to waiting for more information, and some amount of procrastination. Further time was wasted because of lack of access to a computer. What follows represents the beginning of a preliminary NATIONAL roll catalog and some shared knowledge about a fascinating automatic piano company.
I’ve enjoyed looking through this list and hope you will too. It allows us to reach a few conclusions about how NATIONAL conducted its business. Surprisingly, rolls were being released (by someone) as late as 1937!
5000 — 5999 .......1920-21
6000 — 6999 .......1922
7000 — 7999 .......1923-24
8000 — 8100 .......1925
8200 — 8600 .......1926
8601 — 8800 .......1927
8801 — 9440 .......1928-29
9441 — 9629 .......1930
9630 — 9999 .......1931
X-14000 — X-14300 ....1932
X-14301 — X-14450 ....1933
200 — ??? ........1933
X-14451 — X-14660 ....1934
X-14661 — X-14750 ....1935
X-14751 — X-14890.....1936
X-14891 — X-?????-----1937
That’s it. Comments, corrections and suggestions are welcome. If a significant amount of new info comes along, I’ll issue a revised list. Thanks again for the info you sent.
Ed Sprankle
The original rollography report of October, 1985, consisted of eleven single-spaced pages, with a separate two pages for Notes and Keys to National Nickelodeon Roll List. The so-called keys provide an understanding of the comments and/or abbreviations scattered throughout the rollography list. The keys are simple and anything referring to them has been expanded within the new database, so that no special "keys" to understanding the database reports are necessary. The rollography takes into account only minimal recut information, a few terse comments, sparse composer/publisher information, and the number of copies of a particular roll surveyed (if greater than 1) within the scope of the rollography project. Recording the number of rolls for a certain roll title was thought to be possibly useful in determining its popularity, and by inference the relative quantity of rolls cut for that particular title.
Over the intervening years, after the joint National rollography project activity had more or less ceased, John H. Perschbacher continued to carry on, essentially working independently of the earlier work, looking for and compiling additional rollography information. One of the attributes Perschbacher noted, and that was overlooked by the earlier project, was the re-purposing of the top playing note for use in controlling the retrofit Dancing Figure trade stimulator, a detail that has been incorporated into this modernized version of the rollography. But he also made available a nice collection of National tune title strips (for use in the coin slot program window on National coin pianos). These cards provided additional tune information that was happily added to the National database, and in some instances, the tiny cards were the only source of music roll information.
The 65-note (A to C#) music rolls made for the National automatic piano are always, more or less, the length of a single tune music roll. However, continuous medley arrangements incorporating more than one tune have been noted, and rolls with two or three distinctly different selections have been observed and accounted for in the database. But some other roll characteristics remain consistent throughout, with the rolls being invariably 12" wide, and cut with a 6 per inch hole spacing. For the most part the perforations are round, but an early perforator used by the company reportedly cut square holes. It is estimated that about 3,000 musical arrangements were churned out between 1922 and 1930. This would average 7 new music rolls per week, which would almost certainly allow the pianos to live up to the company's motto boldly lettered across the clear glass front, which proclaimed: “Music Changed Weekly.”
But there was another short-lived commercial music roll in the National stable, and it was made from the single tune master rolls. A few of the piano arrangements were punched out on the 9 per inch scale, with standard 11¼” wide paper and boxed under the brand name of “8T8 Har-Mo-Nie (8T8 is a clever abbreviation pronounced eighty-eight). These rolls were intended for the home player piano market. It is likely that the failure of this venture was due, at least in part, to the player piano rolls not being competitively priced compared to the other more established roll manufacturers, who already had extensive distribution networks in place. There is no doubt that some Grand Rapids, Michigan, merchants did stock the “8T8 Har-Mo-Nie” brand, but whatever the case, these 88-note rolls are rare today, with what is probably the largest collection belonging to Willard E. Burkhardt, a long-time collector in Grand Rapids, Michigan. The “8T8 Har-Mo-Nie” brand piano rolls were only produced for a short period of time, from 1920 to 1922. An employee who operated the roll perforating department, during this same era, recalled making some ten-tune coin piano rolls using the National master rolls. Although he did not recall who the customer was, he did remember sending out an order about once a month. Except for these few exceptions, all roll arrangements were only issued for consumption on the National automatic piano routes, and by 1925 there were reportedly 4,000 coin pianos out on location, and then by 1930 there were reportedly another 200, for a total of 4,200 pianos.
In October of 1980, the late Mike Montgomery, who conducted a research study regarding blues roll production, produced a summary of his work. He was a widely known and enthusiastic collector of player piano music rolls, along with being a talented pianist and an avid researcher. Blues music was high up on his lists of interests, and when he learned that the National Piano Manufacturing Company had issued music rolls of blues music composed by George W. Thomas, Jr. (1883 – 1937) he immediately became quite interested in collecting National piano rolls featuring blues arrangements. During the development of Mike’s research he came to know fellow musician Sippie Wallace (born Beulah Belle Thomas, 1898 –1986), the sister of the blues composer George W. Thomas. Sippie Wallace was a musical personality in her own right, and loaned Mike many historic documents she had kept relating to her brother. This treasure trove of aging paper records included many royalty agreements made with the National Piano Manufacturing Company of Grand Rapids, Michigan. The composer George W. Thomas did business as George Thomas and Company, and he is said to have been involved in, or perhaps even the owner of the Chicago Music Publishing Company.
From his extensive research, Mike concluded that National/AMI made music rolls for nearly all the titles for which they had royalty agreements. From the quantities of rolls produced, a minimum run was 13 to 15 rolls, a double run was 30, and so on. From the frequency (or infrequency) of titles they paid royalties on, it seems as though National needed only about 450 to 550 of each blues title, suggesting that there were a series of locations in certain segments of the country that featured blues rolls. National/AMI made each roll in enough quantity to satisfy this “juke box” trade, then went on to other titles. No one title seems to have enjoyed significantly more popularity than any of the others.
Whether the quantity of blues rolls shown in the summary report represents the actual number of rolls produced for other non-blues titles is unknown, but if it is at least somewhat representative it hints at the possibility that National did not necessarily and immediately destroy first-run music rolls when they were removed from a piano. Instead, with some 4,000 pianos to service, each of which needed the music changed weekly, perhaps the latest used rolls were kept busy, circulated amongst other pianos in nearby and/or even distant locations. In effect, then, all pianos could get the advertised weekly change of music, all of it new to any particular piano, but at least some of that “new” music would have been previously used elsewhere. This scenario meant rotating used rolls between various pianos while the tune remained popular, so that each piano had its “new” music program on schedule, while squeezing out every possible nickel for each tune title before retiring it.
Anyone desiring to examine a few documents (in PDF format) relating to Mike Montgomery’s research on blues rolls is invited to choose from the following list:
Before and during the years the National Automatic Piano Rollography project was active, circa 1984-1985, there were several National roll recutting projects either known to have taken place or that were currently active. There may have been more, but the following list gives the name of those people making National recut rolls available:
Although Ed Sprankle reported 508 music rolls in his rollography report of October 7, 1985, after consolidation of its contents during database compilation the total comes to only 482 rolls—26 less than originally reported. This relatively small difference resulted once duplications were eliminated, certain corrections were applied, when multi-tune rolls were counted as a single roll, and when recut roll entries were not counted when they referred to an actual music roll already in the database. Even with the aforementioned adjustments, all pertinent information remains in the database, although perhaps expressed differently than it was in the original Sprankle rollography listing.
But the pure Sprankle rollography was even more transformed once John Perschbacher’s National roll list was overlaid, a process that added bits and pieces of information to the previous compilation, bringing about some welcome improvements to the database records. Moreover, the Perschbacher list was useful in corroborating the accuracy of many Sprankle rollography items. But accommodating change and evolution is the reason for having a flexible and easily modified database, making it easy to add and manipulate the data in useful ways, as well as to allow new data to flow seamlessly into and improve what already exists. And yet, with all this diligent work and effort, it is a certainty that there are still unrecognized errors carried over from the originating data resources, and perhaps more mistakes dared to creep in when the database was being populated. Thus, it is vital that readers make it known when errors are noted, and then contribute updated information whenever possible.
The most accurate source by far is the information provided on the front or trailing leader of a music roll, followed by a National Program or Tune Title Strip, which originally accompanied each music roll. The tune title strips are usually easier and more accurately read than the often battered information stamped with a small font on now dirty and torn roll leaders. Whenever these often lost or misplaced title strips, or a good copy, are available they can be used to verify the rollography information, such as the roll number, and to correct any text deficiencies relative to the tune title and/or composer/publisher information. Thus, the information on music roll leaders and/or the tune title strips are at the top of the trust list for music roll information. Consider the following list suggesting an order of reliability:
Looking back in time, the guidelines for the Sprankle rollography, circa 1984-1985, were much more casual than is the intent behind this new database project. It was never intended to be a formal all-encompassing undertaking. As such, there were many deliberate omissions, such as overlooking certain “unimportant” tune title details and composer/publisher information. Fortunately, some of the previously ignored information has been recovered whenever a suitable alternate source of information has been made available.
—Please send an e-mail message to and include the following information:
Thank you for any assistance you may provide. Information submitted will be added to the National music roll database and/or will be very helpful in confirming that data already collected is correct.
Adobe Reader (or its equivalent) is required to open and/or print the database reports. The free Adobe Reader can be downloaded by clicking on the icon at left.
Download the current database report as a PDF by clicking the left hand button, or report more music rolls by clicking the right hand button. |
29 pages. |
13 pages. |
All database report information is offered "as is," without any guarantee or warranty whatsoever of any kind, neither stated, implied, nor inferred, as to the accuracy, correctness, exactness, suitability, or usefulness of any content.
Ed Sprankle, Willard E. Burkhardt, Claude Garibotti, Mike Montgomery, John Perschbacher, and Charlie Smallwood. New database structure by Terry Hathaway.
Terry Hathaway