Abstract
Some sources of error that arise in the use of Stuckelberg's formula for excitation-transfer calculations are pointed out. To demonstrate these difficulties, Stuckelberg's formula for noncrossing levels is used to calculate nearly adiabatic excitation-transfer cross sections for He*, Rb*, and Cs* collisions with inert gases and for alkali*-alkali collisions. In contradiction with another (erroneous) calculation for He*, these cross sections generally disagree with experimental observations by orders of magnitude. The reasons for this, and some conditions for the validity of this method of calculating nearly adiabatic cross sections are given.