It was shown by A. Neishtadt that dynamical bifurcation, in which the
control parameter is varied with a small but finite speed epsilon, is
characterized by a delay in bifurcation, here denoted lambda(j) and de
pending on epsilon. Here we study dynamical bifurcation, in the framew
ork and with the language of Landau theory of phase transitions, in th
e presence of a Gaussian noise of strength sigma. By numerical experim
ents at fixed epsilon = epsilon(0), we study the dependence of lambda(
j) on sigma for order parameters of dimension less than or equal to 3;
an exact scaling relation satisfied by the equations permits us to ob
tain for this the behavior for general epsilon. We find that in the sm
all-noise regime lambda(j)(sigma) similar or equal to a sigma((-b)), w
hile in the strong-noise regime lambda(j)(sigma) similar or equal to c
e((-d)); we also measure the parameters in these formulas.