Mechanism behind "Hoot Tube" demonstration For years I've been puzzled by how the "Hoot Tube" demonstration shown at the end of this video (around 2 min 45 sec) actually works.
I've done the demo myself with a double walled stovepipe from the hardware store held above a Meker burner.  I did not use the metal mesh shown in the video.  The sound produced is VERY loud and impressive!  You have to hear it in person -- it's a terrific demo.
The only explanations I've encountered are rather vague handwaving arguments about resonance like in the above video.  I'd like to understand what's going on in more detail.  I can testify that the position of the burner head inside the tube is critical, and I was told that an ordinary bunsen burner doesn't work, only Meker burners with the bigger multi-hole head.  Does that somehow mimic the metal mesh with many "holes" in it?
I first saw this demo in a two-week physics workshop by Dick Minnix and Rae Carpenter many years ago at VMI.  Wonderful fun.  I'm still stumped on exactly how it works, though.
 A: I stumbled upon an explanation of the Hoot Tube on Wikipedia.  But until now I've never heard it referred to as a Rijke tube. It looks like this is the mechanism I was looking for --

"The sound comes from a standing wave whose wavelength is about twice
  the length of the tube, giving the fundamental frequency. Lord
  Rayleigh, in his book, gave the correct explanation of how the sound
  is stimulated.[7] The flow of air past the gauze is a combination of
  two motions. There is a uniform upwards motion of the air due to a
  convection current resulting from the gauze heating up the air.
  Superimposed on this is the motion due to the sound wave.
For half the vibration cycle, the air flows into the tube from both
  ends until the pressure reaches a maximum. During the other half
  cycle, the flow of air is outwards until the minimum pressure is
  reached. All air flowing past the gauze is heated to the temperature
  of the gauze and any transfer of heat to the air will increase its
  pressure according to the gas law. As the air flows upwards past the
  gauze most of it will already be hot because it has just come
  downwards past the gauze during the previous half cycle. However, just
  before the pressure maximum, a small quantity of cool air comes into
  contact with the gauze and its pressure is suddenly increased. This
  increases the pressure maximum, so reinforcing the vibration. During
  the other half cycle, when the pressure is decreasing, the air above
  the gauze is forced downwards past the gauze again. Since it is
  already hot, no pressure change due to the gauze takes place, since
  there is no transfer of heat. The sound wave is therefore reinforced
  once every vibration cycle, and it quickly builds up to a very large
  amplitude.
This explains why there is no sound when the flame is heating the
  gauze: all air flowing through the tube is heated by the flame, so
  when it reaches the gauze, it is already hot and no pressure increase
  takes place.
When the gauze is in the upper half of the tube, there is no sound. In
  this case, the cool air brought in from the bottom by the convection
  current reaches the gauze towards the end of the outward vibration
  movement. This is immediately before the pressure minimum, so a sudden
  increase in pressure due to the heat transfer tends to cancel out the
  sound wave instead of reinforcing it."

A: Lord Rayleigh's explanation is classic and is basically correct, but if you research further you'll see allot of debate and other explanations as well. You can create the thermoacoustic resonance by even simpler means than a metal tube and gauze. You don't necessarily need the gauze and can cause the resonance with just a flame and a cardboard tube as I did here:
Also last year I met some students from the University of Connecticut at the American Control Conference. They were interested in the Rijke tube from point of controlling the instability. So to do that they had to model the physics of the tube, and they came up with some interesting results that predict if and when resonance will occur. It turns out to be a very nonlinear map. You can read more about it and research the papers on the website here if you are truly interested.
