Abstract
Since its introduction in 1966, an empirical quadratic fit to Baraff's curves due to Crowell and Sze, has been widely used, in conjunction with the use of Baraff's theory to extrapolate experimental data for electron and hole ionization coefficients to conditions-notably temperature-differing from those under which those data were derived. The quality of that quadratic fit is reexamined, and an improved cubic fit is proposed.