Originally Posted by
Fly-by-Wife
Complete tosh. The molecules of metal in the engine are not free to move about [...] Therefore there is no lowering of density, and no heat rising in the metal of the engine.
To be fair, by heating the bottom of the engine, the outside air around it will presumably receive some heat too and then rise, transferring some heat to the top of the engine by convection.
Whether the amount of heat transferred this way is significant, compared to the amount of heat transferred internally by conduction, is debatable, but could be enough to make bottom-up heating preferable to top-down.