Probably something a bit long, pointy and rather quick...
When trying to identify a fuel gauging problem on another aircraft (loss of fuel quantity at the top of the climb), I did a lot of tests with a fuel gauge in a vertical perspex cylinder of Jet A1. Having it connected to an air data test set, I was able to simulate various rates of climb with varying fuel levels.
Fresh fuel would start to significantly de-gas at about 16,000ft, and over about 20 minutes would trail off to nothing. At around 20,000ft, bubbles were 1/2" diameter as they got to the top of the cylinder. Imagine opening a two litre bottle of sparkling water on a warm day and you will get the idea.
I used the fuel test facility built for the four-engined aeroplane mentioned, to do some fuel tank inerting research much later on. At the time, we thought that removing the dissolved oxygen in the fuel would be important, and tried all sorts of aquarium diffusers to put fine streams of nitrogen enriched air (AKA not quite pure nitrogen) through the fuel, although this proved to be unnecessary in the long run, as the dissolved oxygen is quite small in comparison to the ullage (space above the fuel) on anything other than a tank which is completely full.
Fuel degassing can be a problem in fuel lines which have a valve at each end, such as a tail to wing tank. If the valve are only opened at altitude, the fuel in the pipe will then de-gas, making a gravity transfer more difficult.
Piston-engined aircraft burning petrol type fuels are a different matter. We we told that the B-29 lost about 15% of its potential range due to fuel evaporation, although I can't find a reference at the moment to back that up.
Last edited by Mechta; 11th Jun 2017 at 12:07.