(Phil) The instrument does get a tiny bit shorter (about 2 mm at -50°C) but the overwhelming effect is the reduction in the speed of sound as the temperature goes down. Since there is a standing wave in the instrument the lower speed causes it to take longer to go down and up the instrument, thus lowering the frequency and therefore the pitch. In effect the instrument becomes longer as far as the standing wave sees it. Of course the player's breath warms it up quite a bit, about halfway beteen ambient and body temperature. Even allowing for this (which I didn't in my original posting), the adjustment is about one inch on the tuning slide for 20 degrees and this is just a bit more than is usually available. So if you're playing outside at 0°C all you could do would be to play "short", as they say. Fortunately I don't do marching bands or Sally Army stuff.

So presumably in old, draughty churches without heating it must be impossible to get accurate tuning for the organ for the same reason.

It's pretty difficult to get accurate tuning on any kind of church organ in any case...

True. Accurate was the wrong word... consistent?

(Darren) That's correct. The whole instrument would be flat and it wouldn't sound wrong but any other instrument accompanying it would have to tune down a bit. Rather difficult with a piano. One way round it would be to have a small bleed of a lighter gas into the air the organ uses. You could use hydrogen, helium, neon, methane, ammonia or hydrogen fluoride. Perhaps I should patent this lunacy.