I'd call this a bug: (%i10) sqrt(x^3) * (1/sqrt(x^3)),domain : real, radexpand : true; (%o10) sqrt(x^3)/x^(3/2) (%i11) sqrt(x^3) * (1/sqrt(x^3)), domain : real, radexpand : all; (%o11) 1 (%i12) sqrt(x^3) * (1/sqrt(x^3)), domain : complex; (%o12) 1 --Barton