The full solution of the gas dynamic equations coupled with the
equation
determining the motion of the bubble wall is presented. The
conclusion is that for most of the time the adiabatic approach
(i.e., the assumption of spatial uniformity of the thermodynamic variables)
is a very good approximation. But, for the short time the bubble spends
close to the minimum radius, the full solution of the gas dynamic equations has to
be used. The energy loss and change of equation of state of the gas
resulting from the high temperatures and pressures in the collapsing
bubble are included. A quantitative explanation of the effect
of sonoluminescence is given which is in reasonable agreement with experiment.