The consistency of the Adler-Weisberger sum rule in the large-N expansion is examined. It is shown that the Δ saturates the sum rule in the nonrelativistic quark model to all orders in [Formula: see text] and in the Skyrme model to leading order in [Formula: see text]. Phenomenologically, it is evident that either gA or the integral over the difference of πp total cross sections appearing in the sum rule is poorly behaved as an expansion in [Formula: see text]. In the Skyrme model, based on the calculation of the Δ contribution to the sum rule, it appears that it is gA itself which has large higher-order corrections. It seems likely that in a direct calculation of gA, it it is necessary to include the first two subleading orders before an accurate result for gA can be obtained.