Karl Ellison muses on choosing rolls at a distance and says he tries to
divine them.

He should take the course I do on dowsing, which is the finding of
things, or the finding out of things, using the intuitive sense and,
usually, some kind of tool such as a pendulum.

This is a big subject so I'll stick to piano rolls. I can usually date a
roll to the nearest year by dowsing - this faculty is checkable when
doing pianos themselves using a Piano Atlas - and quite often get the
production run quantity, though this can't be checked except with recuts.
The master of the Rachmaninov Variations on a Theme of Chopin, a 3-roll
set of wonderful music done by Aeolian in London in 1921, seems only to
have been put through twice, which means 32 sets total. All the rolls
I've seen of this have the same strange striations in the paper, so this
seems quite likely.

With a catalog or auction list, I first mark up any rolls that give a
MUST GET indication, regardless of what the print says. Then I go through
the list more conventionally allowing my prejudices full play and when I
come to something promising, get a rating using the dowsing sense. In
hindsight I would say that this process is about 85% successful in
producing a shipment of enjoyable rolls, but of course I have no idea of
the good rolls it's rejected. Quite often the not-enjoyable ones turn out
later to be very rare examples worth keeping for their curiosity value,
though my house is not in a state to support many of such luxuries. The
system is quite excellent for picking up obscure and beautiful salon
pieces.

I can't really stand J. Lawrence Cook because he's such a faker, though
the transition from genuine jazzman to hack in the late 1920s is
interesting, love early Max Kortlander and think Jimmy Blythe is a great
"undiscovered" master.

Sometimes titles can be misleading.  Capitol 'C' roll recut CHICAGO STOMP
(included in a Chicago AMICA Convention medley roll by QRS) is actually a
rare example of contemporary mainline boogie-woogie.

Dan Wilson