5.4 (3). Repeat problem 1, but use the numerical derivative Df(r1) from (5.83) and estimate the error using (5.85) with c replaced by r.