The topic of the differing turntable speeds has been fascinating.
I have at least five turntables with the four different speeds on them,
but I have never had a 16 rpm disk, unfortunately, to try out that
speed.

It was Doug's mention of the Caruso recording that really made me join
in the discussion.  Caruso has always been my favourite singer (even in
primary school when all the kids thought I was nuts.  They didn't know
what to say when I brought along my portable HMV phonograph and Caruso
records to school when I was 10 years old...)

Anyhow, the speeds on Caruso's disks vary tremendously.  His first
recordings on April 11, 1902 were recorded at 71.29 rpm.  His second
session on November 30, 1902 was recorded at 67.92, the first session
at 78[.26] was not until February 1, 1904.  But!  On April 8, 1904 he
recorded the "Mattinata" (with the composer, Leoncavallo, at the piano)
at 73.47 rpm; then the speed seems to stay at around 76.6 rpm for four
years until 1908 when he recorded in January 1910 at 80 rpm, then the
speed jumps back down to 78 for one session (February 7, 1908), and
back down to 76.6 and often 75 for the rest of his recording career,
until September 16, 1920!

In the 1920s and 30s, attempts were made to re-record Caruso's
disks with electrically recorded orchestral accompaniments.  The seven
sides issued by Victor seem to have worked quite well, but when they
abandoned the idea in the midst of the great depression, HMV at Hayes,
Middlesex took over with far worse results.  In fact sometimes the
synchronization of tenor and orchestra is so bad the records are
laughable.  I use these disks to save wear on my precious originals
(I try to ignore the accompaniment).  As I have been listening to
Caruso records since I was five, I wish I had the chance to conduct
an orchestra on this sort of overdubbing - I know every note Caruso
sang inside and out.

The 1909 recording of the Flower Song from "Carmen" (Il fior) was
re-recorded on October 23, 1936 at the HMV studios with Lawrence
Collingwood conducting.  Perhaps this accounts for the 60 Hz hum that
Doug can hear on his copy? This could be identified with the matrix
numbers: the original 1909 record had the matrix number C-8349-1, and
the new electrical take had the matrix number of 2EA 4093-2A.  In
Australia (where I live) it appeared as catalogue number DB 3023, or
DB5388 (depending on the coupling).

I hope this information is not boring, and please accept my best
wishes,

Glenn Amer

(Currently enjoying the latest addition to my collection, an
outstanding original condition Hammond H-112 with all its original
books and receipts, a gem of a lovely organ - and almost given to me!)