Rather than choose an arbitrary partition, divide [a,b] into n equal subintervals [x0,x1],[x1,x2],…,[xn−1,xn], so that
x1−x0=x2−x1=⋯=xn−xn−1=nb−a.
Put h=nb−a; then xi=a+ih, i=1,2,…,n.
Deriving the limit formula. Using the right-end rule (ξi=xi) from §9.2.1,
∫abf(x)dx=limn→∞∑i=1nf(xi)(xi−xi−1)=limn→∞nb−a∑i=1nf(a+i⋅nb−a),
which (renaming the summation index i→r) gives the standard working formula
∫abf(x)dx=n→∞limnb−ar=1∑nf(a+n(b−a)r).
A short computation (isolating the r=0 term, which is nb−af(a)→0 as n→∞) shows the sum may equivalently start at r=0: n→∞limnb−ar=0∑nf(a+n(b−a)r) gives the same value.
Special case a=0, b=1. The formula collapses to
∫01f(x)dx=limn→∞n1∑r=1nf(nr). …