We consider the response of a dynamical system driven by external adia
batic fluctuations. Based on the 'adiabatic following approximation' w
e have made a systematic separation of timescales to carry out an expa
nsion in alpha\mu\(-1), where alpha is the strength of fluctuations an
d \mu\ is the damping rate. We show that the probability distribution
functions obey the differential equations of motion which contain thir
d-order terms (beyond the usual Fokker-Planck terms) leading to non-Ga
ussian noise. The problem of adiabatic fluctuations in velocity space
which is the counterpart of Brownian motion for fast fluctuations, has
been solved exactly. The characteristic function and the associated p
robability distribution function are shown to be of stable form. The l
inear dissipation leads to a steady state which is stable and the vari
ances and higher moments are shown to be finite.