Casual observation shows that the aurora, along a vertical axis, have a color gradient similar to natural chromatic dispersion. That would be expected if the aurora were reflections of light from the central sun, shining through the polar hole, and a mere coincidence if the aurora were from excitation of charged particles.
The chromatic layers in aurora could be explained as the inverse of mirages from the sun during sunset, that light up ice-crystal sheets in the atmosphere that follow the magnetic field lines. Water vapour from the "polar fountain" (NASA, 1997) would reach up into the thermosphere and form sheet-like "clouds" of ice-particles, similar to the "plume" of water vapour at the pole of Europa that reaches 200 km up in the atmosphere (NASA, 2012). Those clouds would follow the magnetic field lines, similar to what Kellner described in 1963.
Synapses
Earth weaves its own invisible cloak - Polar fountains fill magnetosphere with ions - NASA (1997)
Hubble Space Telescope Sees Evidence of Water Vapor Venting off Jupiter Moon - NASA (2012)
Alignment of Cirrus Clouds along the Magnetic Meridian - Nature (1963)
Expansion tectonics and the hollow Earth theory
On the idea that planets are mostly hollow, for example the observation that the "moon rang like a bell" after crashing sond into it (read history, it is well documented), based on the theory of radial expansion (in contrast to plate tectonics that posits a fixed radius), it is quite intuitive that it would be hollow, since if you reverse continental drift and also the continental shelf you see a 4-fold change in radius, that is a 50-fold change in volume. One way to grow that way, without increase in mass, is if the mass is pushed outwards and surrounds a hollow interior.
Here is a simple experiment with a rotating sphere of water in microgravity (close to zero gravity), filled with air bubbles, showing how distribution of matter is affected by rotation.
A model for seismology on a hollow Earth
This model for how seismic waves from Earthquakes spread, that they loose energy with distance, then converge at other side of Earth, 180°, on a globular hollow planet with a lithosphere crust, with a "shadow zone" between around 104° and 140°, fits perfectly with the data.
This is Titan in infrared at 5 micron taken in 2009[1], could this be a polar hole, on a hollow Titan, and the thick atmosphere of Titan reflecting light on those wavelengths from the internal "sun" (a plasmoid) just like you clearly see the atmosphere of Titan does around the rest of the moon as those wavelengths are reflected from our sun?
The x-rays emitting from Jupiter's poles that CHANDRA discovered two decades ago[2], could they be signatures of polar holes on Jupiter, a hollow gas giant?
