A recently developed criterion for A-segregation in steel ingots and castings, based on a dimensionless Rayleigh number, is applied to case studies of two steel ingots. In both cases, experimentally obtained locations of A-segregation are correlated against Rayleigh numbers predicted through casting simulations. It is found that the value of 17±8 previously reported as the critical value for the Rayleigh criterion agrees well with the results of these case studies, but that 6 is an appropriate critical value if a conservative lower bound is sought. Limitations and sources of sensitivity in the application of the criterion are highlighted.