Recently I had a breakthrough in creating an index of the photocopies
of all the known Wurlitzer APP catalogs.  This was finding an old OCR
program that came with my Apple flat-bed scanner.  I had to reload
the scanner driver and found the manual for the OCR program.  While
quaint, the program was intended for converting "on-line" fax
documents.  In other words, it had the feature I needed which was to
allow me to draw a box around some text and convert just that text.
This program will almost read clear, sharply focused text in any
photograph.

Currently there are 5,338 song titles listed in the combined database.
The master index is made from the following catalogs and roll lists:

(1) Wurlitzer APP Catalog #7, last roll 20695  -- This catalog was
loaned to me by Dana Johnson and the high resolution 300 dots per inch
scans were made from this catalog.

(2) Wurlitzer APP Catalog #8 -- This catalog contains selections from
Ziegfeld Follies of 1925.

These two catalogs are the only primary source documents.  Spot checks
of the song title dates indicate that these catalogs date to about the
1926-1927 time frame.  The last roll no. is 20787.  A photocopy of this
catalog was loaned to me by Matthew Caulfield.

The next oldest catalog was produced by Player Piano Company of Wichita,
Kansas, dated November 1969.  On the back is written "Prices slashed up
to 40%."  10-tune rolls were reduced from $22.50 to $15.75, 5-tune
changer rolls were reduced from $13.75 to $8.25.  There is a date on
the corner of this addenda of 1/1/70 (I was only 10 years old at the
time).  A photocopy of this catalog was loaned to me By Matthew
Caulfield.

The roll numbers used in this catalog are somewhat random.  Where they
overlap the 1920s catalogs, the songs are not the same ones.  It appears 
that the rolls in this catalog were made by choosing "popular" songs
and creating new rolls.  Note the large gap in dates between these
catalogs -- close to 43 years!

The next catalog is from Player Piano Center (Vancouver, BC, Canada)
circa 1975, the so-called Doyle Lane catalog.  I have copies of this
from both Matthew Caulfield and Doug Hershberger.  This catalog is the
most compete and details rolls 30297 to 30336 inclusive, with no gaps.
The masters for these rolls are currently in the possession of the
Herschell Carrousel Factory Museum.  Through recent correspondence over
the last few weeks with Doug, and earlier correspondence with Matthew
Caulfield, roll 30301 can be dated to 4-10-1939 and roll 30336 has
songs that date to early 1967.

As for Wurlitzer Caliola rolls, I have almost all of these rolls
scanned as MIDI files, most from the collection of the late Tom Grace.
This week a MMD reader loaned me some of the rolls I was missing,
so that by the end of the month I will have this complete collection
scanned and converted to MIDI files for preservation and study for
qualified researchers.  Any wide distribution of these MIDI files will
be the decision of the Herschell Carrousel Factory Museum who currently
holds these masters.

The rest of the listings can not be considered a primary source.
Ray Siou issued a broadsheet on 8x14 inch paper, dated 9-21-1987 and
which contains a dedication to Charles Smallwood.  Four of the six
pages of this catalog contain APP roll listings, the other two contain
Caliola roll listings.

I made the mistake of scanning the APP pages at too low resolution (100
dots per inch grayscale) instead of 300 dots per inch.  This is too low
to use the optical character recognition to read the roll numbers from.
Even on the photocopy of this roll loaned by Matthew the roll numbers
were hard to make out.  This is the only catalog which has not been
fully integrated into my index.  Interesting enough, many of the late
1920s to early 1930s rolls that survived are listed in this catalog.

Ray Siou liked to give value my combining a 5-tune roll with a 10-tune
roll to create a 15-tune roll.  While this made good business sense,
it makes constructing the index awkward as 5-tune rolls were often 
made by splitting masters into two from two 10-tune rolls.  According
to the description of Wurlitzer roll production by Matthew Caulfield
at http://www.wurlitzer-rolls.com/ , this could be done by changing
the gearing on the perforators.  Doug Hershberger confirmed this to me
by sending me photocopies of some of the master roll leaders, which
indicate that the same master was used to produce a 10-tune APP roll,
a 5-tune changer roll, and a Caliola roll.  On several of my Caliola
rolls are also written all three roll numbers.

What makes things difficult here for the archivist is that, due to the
lengths of the songs or the popularity of the titles, a 10-tune APP
roll does not split evenly after song #5.  The late rolls produced by
T.R.T. were "Length of Ten Tune Roll", so to make 5-tune rolls one song
was omitted.  Then later a roll was created with 5 of the omitted
songs.

A listing of rolls similar to the one issued by Ray Siou was printed on
8.5" x 11" paper.  I believe this to be the main catalog of Play-Rite
re-cuts.  Several years ago a local collector shared with me MIDI Files
of most of these rolls -- about 2,000 MIDI files, and the inspiration
for creating this index.   What surprised me when I cross-referenced
these MIDI Files with the above catalogs was the discovery that a number
of the better early- to mid-1920s rolls survived, more than the "high
brow" classical I mentioned in an earlier posting about these rolls.

In answer to those postings, Jack Conway and Don Teach made some copies
of collection listings, including Ed Sprankle's handwritten listing of
rolls.  Because of the overlap of the printed catalog I was able to enter
the rest of the roll titles in by hand last week into the master index.

Also included in the master index are some lists Matthew Caulfield kept
in a folder, were primarily the listings from Hal Davis.  In addition
to this and notes from Frank Himpsl there were copies of eBay listings
which are useful in cross-referencing back to the other roll catalogs.
These have been included in the master list.

Still to be added into the master index are the 5-tune rolls from Jack
Conway and Don Teach.  This is a bit confusing as the roll numbers have
been changed: Wurlitzer numbered the 1920s 5-tune rolls with a 4-digit
number starting with 2.  When I borrowed Matthew Caulfield's Catalog
#8 I only scanned the 5-tune roll index at 72 dots per inch grayscale,
I did not scan the main catalog at all.  As far as I can tell the songs
are the same.  I regret now that I did not retain and copy the full
catalog as I now know that the 5-tune rolls are NOT made by splitting
a 10-tune roll in half at tune 5.

Ray Siou and Play-Rite renumbered the rolls to use a number between
1 and about 118.  Often the catalog gives the Wurlitzer number in tiny
type.  It is unclear when this practice began.  T.R.T. used the last
digits of the APP roll number followed by a dash (-) and then a digit
1 or 2.  Most of these seem to split the roll at tune 5.   There are
just enough cases of mixing songs from the masters to render this
numbering system useless.

I am still working on cross reference checking the data from the
different catalogs.  I would like to get better copies of Ray Siou's
broadsheets before I post the full index to the MMD.  Currently the
listing is in an Excel spreadsheet, sorted alphabetically by song
title so as to see what songs are duplicates.  It should be possible
with some help to reformat this into a listing similar to the other
roll catalogs archived on Wurlitzer-Roll.com.

Before that is done I would like to get some help in completing the
entry and checking the spelling of the composer names which were
printed on many of the roll box labels.  Since I scanned the index of
the catalog not all of this was captured.  I did back-fill in the roll
track number information into a separate column.

The other help that I would like, would be to look up the copyright
dates on all the song titles.  This would be useful for re-creating the
late 1920s to early 1930s catalog, which seems to have gone missing.
Since there are over 5,000 line items, this is too much for a single
person to do.  The practical way would be to distribute the task of
verifying the data to several people.

Julie Porter