The density of the flow object defines the initial density of the smoke
that's created (from each particle). The new density (previously
gravity) setting of the domain defines how much the current density at
any point of the smoke effects upwards force at that point. Similarly the
flow
Hmm, indeed this is not very logical. The gravity and heat parameters
actually change how much the density and heat of the smoke effect the
motion of the smoke, so they're not directly connected to gravity. I'll
change the parameter name and add a bit more descriptive tooltips to both
of