You have that the error in the polynomial value is less than (M is an upper bound on ) so the error in the integral is less than the integral of that from 0 to 1. And since that is a constant, it is just that number times the length of the interval, which is 1.